# Managing users (/docs/administration/managing-users) # Managing users Content coming soon. # Quickstart (/docs) Nios is an agentic AI platform for value chain intelligence, built for procurement and sustainability teams. Set up your account below to get started. ## Create your account and log in Go to [app.nios.earth](https://app.nios.earth/sign-up). Sign up with your Google or Microsoft account or email and password. Create your organization by providing a name and optionally uploading a logo. After creating your organization, you'll be automatically logged in. ## Next steps Quantify environmental impacts across your upstream value chains. Chat with the agent to analyze your data and calculate impacts for your products. Organize products into projects with files and instructions that guide the agent. ## Need help? * **Support:** Contact our team via [hello@nios.earth](mailto:hello@nios.earth). * **Data security:** Review how we handle your confidential data in our [Data Processing Addendum](/legal/dpa). * **Privacy:** Review our [privacy policy](/legal/privacy) to understand how your personal information is protected. # Introduction (/docs/introduction) import { Icons } from '@/components/icons'; ## About Nios Nios is an AI-powered decision intelligence platform for supply chain resilience and sustainability. Built by and for sustainability and supply chain professionals, it delivers major productivity gains by quickly retrieving the information they typically need in their workflows, automating environmental impact and risk assessments across complex product supply chains, and rapidly providing the deep intelligence needed for confident decision-making. ## Why Nios? Purpose-built for sustainability and supply chain teams, Nios reduces manual effort and the time required to retrieve, process, and analyze data and generate decision-ready outputs. Some highlights include: * **Focus more on outcomes than analyses:** Offload routine modeling and calculations for supply chain environmental impacts, such as GHG emissions and biodiversity, so teams can spend their time interpreting results, deciding actions, and managing outcomes. * **Expert guidance on demand:** Use expert trained chat agents to sense-check assumptions, explore mitigation ideas, and get deep insights instantly. * **Instant access to relevant information:** Pull relevant information instantly from internal documents, environmental datasets, and public sources instead of searching across scattered systems. * **Learn fast:** Let team members understand complex sustainability and supply chain topics through clear, guided chat interactions. # Overview (/docs/overview) ## What is Nios? Nios is an **agentic AI platform for value chain raw material intelligence**, built for **procurement and sustainability teams**. Nios helps enterprises move beyond tier-1 upstream value chain visibility to uncover hidden raw material dependencies, trace likely sourcing geographies, and quantify environmental impacts such as **biodiversity, carbon, water, land, and pollution**, all the way across multi-stage value chains. Today, organizations spend **months on manual analyses, consultants, or static tools** that fail to keep up with fast-changing market environment. Nios transforms this process by combining **customer data, leading environmental databases, and AI agents** to deliver **continuous, decision-ready intelligence in minutes, not months**. ## What is an agentic AI system? An **agentic AI system** is a network of specialized AI agents that collaborate to complete complex tasks. Each agent is powered by a large language model (LLM) such as **Claude Sonnet 4**, wrapped with role-specific instructions ("prompts") and often given access to external tools like databases, web search, or code execution. A central **orchestrator** decides which agent to activate and which model to call (e.g., Claude for reasoning, code-focused LLMs for structured outputs). Agents pass results between each other, so the workflow builds step by step. Think of it like a **project team of experts**: one member researches, another analyzes numbers, another writes up the findings, and the orchestrator coordinates who does what. 👉 **Why it matters:** Unlike a single chatbot, agentic AI can break down complex workflows into coordinated steps, producing results that are more structured, scalable, and closer to real expert work. In short, instead of one model trying to do everything, an agentic AI system works like a **team of LLM-powered specialists**, coordinated to deliver structured results. ⚠️ **Disclaimer:** Outputs are probabilistic and should be interpreted as directional approximations, not exact facts. We are continuously working to increase transparency about how the agents operate, what assumptions and data they use, and the limitations of their results. Think of them as **capable assistants**: understanding what they can and cannot do helps you get the most out of them. 👉 **Privacy note:** Customer data processed by Nios agents is never used to train external models. Inputs are only used to perform the requested calculations and deliver results within the Nios system. # Agent effort (/docs/chat/agent-effort) Nios uses specialized agents for different tasks: creating products, updating bill of materials, comparing impacts, and answering questions. Each agent is optimized for its task with a dedicated intelligence configuration. The **Agent effort** dropdown lets you tune this configuration per message. A higher effort level increases the depth of reasoning the agent applies, regardless of which specialized agent handles your task. ## Effort levels * **Medium (default):** The recommended level for most tasks. Provides the best balance of reasoning quality and response speed for product creation, impact analysis, and standard comparisons. * **Low:** Optimized for speed. Best for simple factual questions and quick data lookups where deep reasoning is not needed. * **High:** Activates the most capable intelligence configuration with extended thinking. Best for breaking down very complex products in bill of materials, or performing detailed value chain comparisons. Expect roughly twice the response time. Start with Medium. Switch to High when you need the agent to reason through complex trade-offs or edge cases, and to Low for quick lookups. Agent effort controls reasoning depth (how hard the agent thinks). [Chat modes](/docs/chat/chat-modes) control behavioral scope (what the agent can do). These are independent controls that work together. # Chat modes (/docs/chat/chat-modes) The chat mode selector in the chat interface lets you switch between two modes that control what the chat can do. Select a mode from the dropdown before sending your message. ## Agent mode Use Agent mode when you want the chat to actively perform tasks for you. In Agent mode, the chat can: * Generate a bill of materials based on your inputs * Update and refine your bill of materials * Compare your calculated environmental impacts with other products Think of it as a hands-on assistant fully integrated in the platform that creates and updates your products. **Example requests** ``` Create a bom for 1 ton of crude steel produced in Germany based on electric arc furnace (EAF) method. ``` ``` Compare the biodiversity impact of switching from conventional cotton to organic cotton. ``` ## Ask mode Use Ask mode when you want to explore or understand your data without changing product data. In Ask mode, the chat can: * Answer clarifying questions about your measurements * Explain methods, results, or data sources * Describe how specific impacts are calculated * Clarify [methodology](/docs/impact-measurement/methodology) choices and confidence levels Ask mode is your safe space for discovery and learning, without altering any data. **Example requests** ``` Why is the biodiversity impact of my wool sweater so high compared to synthetic alternatives? ``` ``` What data sources are used for land use calculations in Southeast Asia? ``` ## Choosing the right mode | Scenario | Mode | | ------------------------------------ | ----- | | Create or update a bill of materials | Agent | | Calculate environmental impacts | Agent | | Run a product comparison | Agent | | Understand what a result means | Ask | | Learn how an impact was calculated | Ask | | Explore methodology or data sources | Ask | # File attachments (/docs/chat/file-attachments) import { PaperclipIcon, PlusIcon } from 'lucide-react'; You can attach files directly to a chat message for immediate analysis. The agent reads the attached files and uses their content to answer your question or perform the requested task. ## How to attach files 1. Drag and drop files into the chat input, or click **** and select ** Add files**. 2. Write your message and send. You can attach multiple files in a single message. For spreadsheets with multiple sheets, the agent can read and reference individual sheets. ### Supported file types | Type | Formats | Max file size | | ------------- | ------------------------------- | ------------- | | Spreadsheets | Excel (.xlsx, .xls), .csv, .tsv | 50 MB | | Documents | PDF (.pdf) | 50 MB | | Documents | Word (.docx) | 20 MB | | Text and data | .txt, .md, .json | 100 MB | Large files are automatically truncated to fit the agent's context window. You'll see an indicator when a file has been truncated. ## How attachments work When you attach a file, the agent extracts its content and includes it directly in your message. This gives the agent full access to the file when responding. The agent can refer back to attached files in follow-up messages within the same conversation. You don't need to reattach a file to ask further questions about it. In longer conversations, the agent prioritizes recent attachments over older ones to stay within its context window. Attachments are scoped to the chat where you send them. They are not added to the project's knowledge base and are not available in other chats. If you need a file available across multiple conversations, upload it as a [project file](/docs/projects/files) instead. ## Preview attached files Click an attachment card in the chat input or conversation to open a preview of the extracted content. The preview shows exactly what the agent sees when processing the file. For spreadsheets with multiple sheets, the preview includes tabs to navigate between sheets. ## Best practices The agent reads the full content of every attached file. The more focused the input, the sharper the analysis: irrelevant content takes up space in the agent's context window and can dilute its attention on what matters. ### Keep files focused Before attaching a file, consider trimming it to the data relevant to your question. For example, if you only need Q4 data from a full-year report, extract that section into a separate file. This helps the agent concentrate on the right information and produce more accurate results. ### Working with spreadsheets The agent extracts all sheets from a spreadsheet automatically. If only certain sheets are relevant to your question, reference the specific sheet in your message so the agent knows where to focus. When creating a bill of materials in [Agent mode](/docs/chat/chat-modes), reference the sheet that contains your product data: ``` Based on the "Raw Materials" sheet in the attached file, create a bill of materials for 1000 units of wool sweaters produced in India. ``` ``` Summarize the key findings from the attached sustainability audit report. ``` ## Attachments vs. project files | | File attachments | Project files | | ------------- | ------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- | | **Scope** | Only the chat the file is attach to. | All chats in a project. | | **Retrieval** | By attaching the file to the chat. | Automatically retrieved based on your message content. [@-mention](/docs/chat/file-references) specific files to prioritize them for retrieval. | | **Context** | Available in follow-up messages. The agent prioritizes recent attachments in longer conversations | Indexed and retrieved fresh for each message from the project files. | | **Best for** | One-off analysis in a specific chat. | Recurring reference data organized by project. | # File references (/docs/chat/file-references) When working in a [project](/docs/projects/overview) chat, you can @-mention specific project files to prioritize them for the agent's response. This ensures the agent focuses on the files most relevant to your question. File references are only available in project chats, since they reference files uploaded to a [project](/docs/projects/files). In standalone chats, use [file attachments](/docs/chat/file-attachments) instead. ## How to reference a file 1. Type `@` in the chat input to open the file picker. 2. Search for and select the file you want to reference. 3. Write your message and send. The referenced file appears as a chip in your message. You can reference multiple files in a single message. Click any reference chip to preview the file content. You can select all files from the current project and its [parent projects](/docs/projects/overview#project-hierarchy). ## How file references work Referenced files are given higher priority in the agent's retrieval. The agent searches referenced files first while still searching other project files with lower priority, so you get focused results without losing broader context. References from recent messages automatically carry over to follow-up messages. You don't need to remention a file to keep it active in the conversation. The agent continues to prioritize it automatically. Carry-over fades naturally as the conversation moves on and new files are referenced. When the agent uses content from a referenced file, it includes a [citation](/docs/chat/sources-and-citations) linking back to the source. ## When to use file references Use @-mentions when you want the agent to focus on specific files rather than relying entirely on automatic retrieval from all project files. **Common scenarios:** * **Large projects with many files:** Steer the agent to the most relevant documents when automatic retrieval might surface less relevant matches. * **Cross-referencing specific files:** Reference multiple files to ask the agent to compare or synthesize information across them. * **Targeted follow-up questions:** Reference a specific report to ask deeper questions about its content. **Example requests** ``` Based on @supplier-audit-2025.pdf, what are the main compliance risks in our cotton supply chain? ``` ``` Compare the emissions data in @factory-report-q3.xlsx with @factory-report-q4.xlsx. ``` ## File references vs. file attachments | | File references (@-mention) | File attachments | | ------------ | -------------------------------------------------------------- | ------------------------------------------------------------------- | | **Source** | Files already uploaded to the project. | Files from your device. | | **Scope** | Available in any chat within the project. | The chat where you send them. | | **Context** | Prioritized for retrieval; carries over to follow-up messages. | Included directly in your message; available in follow-up messages. | | **Best for** | Steering the agent to specific project files. | One-off analysis of a file not in the project. | # Managing chats (/docs/chat/managing-chats) The **Chats** sidebar provides a central view of all your conversations, both standalone chats and project chats. Project chats are visually grouped under their project name, making it easy to distinguish them from standalone conversations. ## Finding chats Use the search field at the top of the sidebar to find past conversations by keyword. ## Filtering chats Filter the chat list to narrow down your conversations: * **By type:** Show all chats, only standalone chats, or only product chats. * **By project:** Select a specific project to see only its chats. ## Moving chats to a project You can move a standalone chat into a project to give it access to the project's files and instructions. This is useful when a conversation evolves from an ad-hoc question into sustained work that belongs in a project. See [Managing projects](/docs/projects/managing-projects) for details on organizing chats within projects. ## Deleting chats Delete a chat from the sidebar when you no longer need it. Deleted chats cannot be recovered. Deleting a project chat removes it from the project. The project itself and its files are not affected. # Overview (/docs/chat/overview) The chat is your agent companion across the entire platform. Use it to create products and bill of materials, analyze environmental impacts, simulate mitigation strategies, and generate reports. ## Starting a chat You can start a chat in two ways: * **Standalone chat:** Start a **New chat** from the sidebar for quick questions or ad-hoc analysis. No project context is attached. * **Project chat:** Start from within a [project](/docs/projects/overview) to inherit the project's files and instructions. The agent automatically draws on this context for every response. ## Key capabilities Every chat gives you access to the following features: Switch between Agent mode (task execution) and Ask mode (exploration). Control the agent's reasoning depth to balance quality and speed. Attach files to a message for immediate, in-context analysis. @-mention project files to prioritize them for the agent's response. Trace the agent's answers back to their sources with inline citations. Find, organize, and manage your conversations. ## Examples of what you can do import { Tabs, Tab } from 'fumadocs-ui/components/tabs'; **Create a bill of materials:** Describe your product to generate a structured bill of materials. ``` Create a bill of materials for my product based on [production method] in [country]. ``` **Compare impacts:** Benchmark your product's environmental footprint. ``` Compare the environmental impacts of my product against industry averages. ``` Recommended mode: **Agent** **Scan risks:** Identify regulatory or environmental risks across your materials. ``` What are the biggest regulatory risks for [material] sourced from [country]? ``` **Benchmark:** Compare your results against alternatives. ``` Compare the biodiversity impact of my [product] against industry averages ``` Recommended mode: **Ask** **Business case:** Estimate the cost-effectiveness of mitigation actions. ``` What would be the cost change and impact reduction if I switched from [current material] to [alternative material]? ``` **Screen alternatives:** Find substitute materials that meet your requirements. ``` Find alternative materials to [material] with lower [impact metric] and similar cost ``` **Evaluate trade-offs:** See how biodiversity, emissions, and costs interact. ``` How would switching to [alternative material] affect biodiversity vs GHG emissions? ``` Recommended mode: **Agent** **Create text extracts:** Generate content for internal and external reporting. ``` Generate a summary of [product]'s environmental footprint for our sustainability report ``` Recommended mode: **Agent** # Sources and citations (/docs/chat/sources-and-citations) The agent cites the sources it relies on, so you can trace claims back to their origin. Inline citations and the sources are available in the chat conversation when the agent searches results from the web or retrieves [project files](/docs/projects/files). They are not generated for [file attachments](/docs/chat/file-attachments). ## How sources are selected The agent automatically searches your project files for the most relevant content when formulating a response. Project files are favored over web results: the agent only searches the web when a question requires additional information not covered by your project files to perform the task. This means your own data takes precedence, and the sources the agent finds directly steer its analysis and conclusions. ## Inline citations When the agent uses information from a source, it places an inline citation badge in the response text. Each badge shows the source name: a filename for project files, or a domain for web results. When multiple sources support the same passage, the badge shows a count indicator you can expand. Hover over a badge to see a preview of the cited content. Click a badge to navigate to the source: * **File citation:** Opens a [file preview](/docs/projects/files#preview-project-files) with the cited passage highlighted and scrolled into view. This lets you verify exactly which part of a document the agent relied on. * **Web citation:** Opens the original page in a new tab. ## Sources panel Each response that uses sources includes a sources panel listing all referenced material, grouped into two categories: * **Files:** Project files the agent retrieved, with the filename and location within the file. * **Web:** Web pages the agent found, with the domain and page title. Click any source in the panel to navigate to it. ## Copying responses with citations When you copy an agent response, citations are automatically converted to numbered footnotes with a "Sources" section appended. This makes it easy to share findings with colleagues while preserving the references. # Code blocks (/docs/essentials/code) ## Inline code To denote a `word` or `phrase` as code, enclose it in backticks (\`). ``` To denote a `word` or `phrase` as code, enclose it in backticks (`). ``` ## Code blocks Use [fenced code blocks](https://www.markdownguide.org/extended-syntax/#fenced-code-blocks) by enclosing code in three backticks and follow the leading ticks with the programming language of your snippet to get syntax highlighting. Optionally, you can also write the name of your code after the programming language. ```java HelloWorld.java class HelloWorld { public static void main(String[] args) { System.out.println("Hello, World!"); } } ``` ````md ```java HelloWorld.java class HelloWorld { public static void main(String[] args) { System.out.println("Hello, World!"); } } ``` ```` # Images and embeds (/docs/essentials/images) ## Image ### Using Markdown The [markdown syntax](https://www.markdownguide.org/basic-syntax/#images) lets you add images using the following code ```md ![title](/path/image.jpg) ``` Note that the image file size must be less than 5MB. Otherwise, we recommend hosting on a service like [Cloudinary](https://cloudinary.com/) or [S3](https://aws.amazon.com/s3/). You can then use that URL and embed. ### Using embeds To get more customizability with images, you can also use [embeds](/writing-content/embed) to add images ```html ``` ## Embeds and HTML elements ``` # Markdown syntax (/docs/essentials/markdown) ## Titles Best used for section headers. ```md ## Titles ``` ### Subtitles Best used for subsection headers. ```md ### Subtitles ``` Each **title** and **subtitle** creates an anchor and also shows up on the table of contents on the right. ## Text formatting We support most markdown formatting. Simply add `**`, `_`, or `~` around text to format it. | Style | How to write it | Result | | ------------- | ----------------- | ----------------- | | Bold | `**bold**` | **bold** | | Italic | `_italic_` | *italic* | | Strikethrough | `~strikethrough~` | ~~strikethrough~~ | You can combine these. For example, write `**_bold and italic_**` to get ***bold and italic*** text. You need to use HTML to write superscript and subscript text. That is, add `` or `` around your text. | Text Size | How to write it | Result | | ----------- | ------------------------ | ---------------------- | | Superscript | `superscript` | superscript | | Subscript | `subscript` | subscript | ## Linking to pages You can add a link by wrapping text in `[]()`. You would write `[link to google](https://google.com)` to [link to google](https://google.com). Links to pages in your docs need to be root-relative. Basically, you should include the entire folder path. For example, `[link to text](/writing-content/text)` links to the page "Text" in our components section. Relative links like `[link to text](../text)` will open slower because we cannot optimize them as easily. ## Blockquotes ### Singleline To create a blockquote, add a `>` in front of a paragraph. > Dorothy followed her through many of the beautiful rooms in her castle. ```md > Dorothy followed her through many of the beautiful rooms in her castle. ``` ### Multiline > Dorothy followed her through many of the beautiful rooms in her castle. > > The Witch bade her clean the pots and kettles and sweep the floor and keep the fire fed with wood. ```md > Dorothy followed her through many of the beautiful rooms in her castle. > > The Witch bade her clean the pots and kettles and sweep the floor and keep the fire fed with wood. ``` # Product benchmarking (/docs/guides/product-benchmarking) # Product benchmarking Content coming soon. # Background measurements (/docs/impact-measurement/background-measurements) # Background measurements Content coming soon. # Bill of materials (/docs/impact-measurement/bill-of-materials) import { PaperclipIcon, CopyIcon } from 'lucide-react'; A **bill of materials (BoM)** is a structured list of all components and materials required to manufacture a product. ## 1. Start the analysis You can create a bill of materials in two ways: Describe your product in the chat interface. Upload a spreadsheet with product data. ### Chat message 1. In the chat, select [Agent mode](/docs/chat/chat-modes#agent-mode). 2. Write a message requesting a bill of materials by describing your product in detail. **Example request** ``` Create a bom for 1 ton of crude steel produced in Germany based on electric arc furnace (EAF) method. ``` The more details you provide, the more accurate the results. **What to include:** * **Quantity:** Amount to analyze (e.g., "1 ton of crude steel") * **Country of production:** Manufacturing location (e.g., "produced in Germany") * **Production method:** How it's made (e.g., "electric arc furnace (EAF) method") * **Specific product or brand:** If applicable ### Spreadsheet upload If you have an existing bill of materials or similar product data as a spreadsheet, you can [attach it directly to your chat message](/docs/chat/file-attachments): 1. In the chat, select [Agent mode](/docs/chat/chat-modes#agent-mode). 2. Upload your spreadsheet by clicking the icon. Excel and CSV formats are supported, including multi-sheet workbooks. 3. Write a message requesting a bill of materials and providing instructions on how to read the uploaded data. **Example request** ``` Create a bom for 1000 units of wool sweaters produced in India. Each row in the spreadsheet represents a material in the supply chain with its prices and quantities. Infer the most likely sourcing countries based on typical global trade patterns and industry practices. ``` ## 2. Review and refine your BoM After you submit your message, the system generates a BoM table that outlines your product's direct components and materials, along with estimates of how these raw materials transform across upstream value-chain stages. The agent also provides transparency by showing its reasoning steps, [citing the sources](/docs/chat/sources-and-citations) it relied on, and explaining why it arrived at each result. The more guidance users provide about the level of transparency they need (or which data sources should be included or excluded), the more precisely the agent can tailor its analysis. Explanation of the columns in the bill of materials table below: | Column | Description | Example | | --------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------- | | **Material** | The component or material used in the product, shown in a hierarchical structure:
1. End product
2. Material manufacturing
3. Intermediate materials processing
4. Raw material production | "SMD Capacitors (ceramic)" listed under the "Main PCB Assembly" | | **Matched Reference** | The closest matching material category or dataset reference used to link the item with the correct environmental impact intensity factors for impact measurement. | Mapped to: "Other non-metallic mineral products" | | **Country** | The sourcing or manufacturing country for the material either based on confirmed primary data, or a likely location inferred based on customer data, Nios datasets, or publicly available sources. | Japan | | **Spend** | Spend-based values estimate the manufacturing cost of producing the material (excluding added margins such as VAT or other markups) and are used in impact measurement to match materials with spend-based intensity factors. | €6.4 total cost for 80 capacitors | | **Base Price** | The per-unit spend of the material or component, aligned with the producer-level price used in spend calculations. | €0.08 per capacitor | | **Quantity** | The number of units needed to produce the product. | 80 units | | **Unit** | The measurement unit associated with the quantity (e.g., units, kg, m). | units | You can refine your bill of materials in two ways: Request updates using the chat in agent mode. Change values directly in the table. ### Use chat 1. In the chat, select [Agent mode](/docs/chat/chat-modes#agent-mode). 2. Write a message describing the changes you want to make. **Example requests** * Change buttons, core labels and sewing origin to China. * Change unit cost of 500kg of wool to 14€. * Add raw materials also for buttons, labels and thread. * Update cotton sourcing from India to Bangladesh. * Include packaging materials: 50 units of cardboard boxes from Germany. ### Edit table * Select a table cell and edit its value. We automatically save the change. * Remove or add rows to refine your bill of materials. ## 3. Calculate impacts After finalizing your bill of materials table, click the **Calculate Impact** button to start the calculation. If you work on the same product regularly, consider uploading your supplier data as [project files](/docs/projects/files). The agent automatically retrieves relevant context from your project files based on each chat message. You can also [@-mention specific files](/docs/chat/file-references) to prioritize them. # Exporting data (/docs/impact-measurement/exporting-data) import { CopyIcon } from 'lucide-react'; ## How to export data You can export both your bill of materials and your impact calculation results using a similar, simple workflow: 1. **Navigate to the relevant view.** * Bill of Materials view for bill of materials * Material View for impact results 2. **Click the button above the table to copy all data to your clipboard.** 3. **Open a spreadsheet application (e.g., Excel) and paste the data.** 4. **Save the file using the recommended naming conventions below.** **BoM exports** ``` ProductName_BillOfMaterials_Date.xlsx Examples: - WoolSweater_BillOfMaterials_2025-10-16.xlsx - AluminumFrame_BillOfMaterials_2025-10-16.xlsx ``` **Impact results exports** ``` ProductName_ImpactResults_Date.xlsx Examples: - WoolSweater_ImpactResults_2025-10-16.xlsx - AluminumFrame_ImpactResults_2025-10-16.xlsx ``` ## Copy agent responses You can also ask the agent in [chat](/docs/chat/overview) to generate text summaries or report extracts suitable for sustainability reports or internal briefings. Copy the agent's analysis and reasoning directly from any chat message using the **Copy** button. **Example requests** ``` Generate a summary of our wool sweater's environmental footprint for our sustainability report. ``` ``` Write an executive briefing on the biodiversity hotspots in our product's supply chain. ``` # Impact (/docs/impact-measurement/impact) import { ArrowLeftRight,Filter, Expand, ChevronDown, Copy, Earth, ThermometerSun, Trees, Dam, Droplets, Package, Layers, Anvil, TreePine } from 'lucide-react'; ## Understand your results The impact results can be analyzed with two complementary views: High-level insights into your products' total environmental impacts. Granular, material-level data with detailed impact breakdowns. **Tips to help you navigate the interface:** * **Bill of Materials Impact:** Toggle in the top-left corner to switch between your bill of materials table and the impact results view. * **Overview Material View:** Toggle in the top-left of the Impact page to inspect results in both high-level summary view and in the detailed material view. * **Tooltips:** Additional information is available throughout the interface in tooltips found via hover interactions. ## Overview The Overview page provides high-level insights into your products' total environmental impacts: **Total impacts** At the top of the page, you'll see aggregate metrics: * **Biodiversity:** Measured in PDF·yr (Potentially Disappeared Fraction of Species over time) * **GHG:** Total greenhouse gas emissions in kg CO₂e (carbon dioxide equivalents) * **Land use:** Total land area used in m² (square meters of land transformed) * **Water use:** Total bluewater consumed in m³ H₂Oe (cubic meters of water-equivalent) * **Water stress:** Water consumption adjusted for regional scarcity in m³ H₂Oe (cubic meters of water-equivalent) **PDF (Potentially Disappeared Fraction of Species)** is an emerging biodiversity footprint metric that estimates the fraction of species globally that are committed to extinction due to nature loss pressures (e.g., land use, climate change, water use, pollution). See more in [Methodology](methodology) section. **Impact breakdowns** Below the totals, you can navigate through the dropdown menu several detailed breakdown views that help you understand where and how impacts occur across categories, materials, and geographies: **Impact Categories:** Distribution by impact categories * **Biodiversity:** By nature loss pressures * **Climate change:** Aggregated greenhouse gas (GHG) emissions * **Land use:** By different types of land use * **Water use:** By agricultural and non-agricultural bluewater use * **Water stress:** By agriculture and non-agriculture water stress See more in [Impact group coverage](methodology#impact-group-coverage) *Use this view to identify which pressures contribute most to overall impact.* **Impact Categories across Value Chain:** Distribution by impact categories across value chain stages * Stage 1: End-product * Stage 2: Material manufacturing * Stage 3: Intermediate material processing * Stage 4: Raw material production See more in [Value chain stage coverage](methodology#value-chain-stage-coverage) *Helps identify where in the value chain most impacts occur.* **Top Materials by Impact:** Materials with the highest impact values * Highlights materials with the largest environmental footprint. * Shows sourcing country and relative impact value. *Useful for prioritizing material-level mitigation actions.* **Top Countries by Impact:** Countries with the highest impact values * Displays which sourcing countries contribute most to total impact. * Supports optimization of procurement and supplier selection. *Helps locate geographic hotspots in your value chain.* **Geographic Impact Distribution:** Impact by country location * Darker shades indicate higher impact levels. * Hover over a country to see specific values. *Provides an overview of global impact concentration of all materials across the value chain.* ## Material view The Material view provides granular, material-level data for all items in your BoM. **Table columns** The table includes the following columns: * **Material:** Each material from your BoM * **Country:** Sourcing location * **Spend:** Total spend per material (quantity × base price) * **Base Price:** Unit price of each material * **Quantity and unit:** Amount of material used * **Environmental impact metrics:** Impacts for each impact group and category **View controls** Use the toolbar above the table to refine and navigate data: * **Filter impact measurement:** Use the search field to flexibly filter results based on your needs, such as by material, country, or impact category. * **Impact type dropdown:** Select from dropdown to switch between aggregated impacts and stage-specific views. * **Categories toggle:** Click to show or hide impact categories. * **Expand/Collapse all:** Click to expand or collapse all material hierarchies. * **Copy:** Select to copy the current table view, for example to export to spreadsheet. See [Methodology](methodology) to understand more about the underlying scientific framework and data used. ## Explore results with the agent You can use the agent in chat to dig deeper into your impact results: * **[Ask mode](/docs/chat/chat-modes#ask-mode):** Ask questions about what the numbers mean, how a specific impact was calculated, or what drives a particular hotspot. Ask mode is a safe space for exploration without modifying any data. * **[Agent mode](/docs/chat/chat-modes#agent-mode):** Create and update your products' bill of materials, and compare your product impacts. The agent's responses about impact results include [citations](/docs/chat/sources-and-citations) back to the underlying data sources, so you can verify claims and trace information to specific documents or databases. # Methodology (/docs/impact-measurement/methodology) The following sections explain the methodological foundations behind Nios impact measurements. You can also use [Ask mode](/docs/chat/chat-modes#ask-mode) in chat to ask the agent about specific methodological choices, how a particular impact was calculated, or the confidence level of your results. Two-step calculation process applied. Value chain stages and impact groups covered. Underlying databases and scientific frameworks used. Explanation of how the Nios methodology is aligned to global standards. Explanation of methodology limitations and planned work. ## How impact calculation works The Nios system applies a two-step calculation process: ### 1. Midpoint environmental impact measurement Each material in your BoM is analyzed to calculate its environmental footprint: * **The Nios agent** maps each BoM material to a relevant **reference material impact intensity factor** from our [environmental intensity factor databases](#reference-data). * The material spend value is multiplied by the intensity factor to calculate environmental impacts in a **standardized and comparable way**. * The Nios agent evaluates how closely the reference material matches your BoM entry, providing transparency in the BoM tables on underlying assumptions. * The system calculates **midpoint environmental impacts**, including **GHG emissions**, **land use**, **water use**, and **pollution**. ### 2. Biodiversity translation Midpoint environmental impacts are characterized into a single endpoint biodiversity impact metric: * **GHG emissions**, **land use**, **water use**, and **pollution** are converted into **PDF·yr** (Potentially Disappeared Fraction of species over time). * Each environmental impact is multiplied by **region- and impact-specific characterization factors** from the **LC-Impact methodology**. * These factors capture how strongly each pressure affects biodiversity in different locations, making results spatially explicit. PDF·yr is a risk indicator or a "common currency" for comparing biodiversity loss risk across pressures, regions, and products. It is not recommended to be interpreted as literal numbers of species lost. ## Key notes on analysis scope ### Value chain stage coverage The assessment scope covers upstream value chain stages, resulting in cradle-to-gate system boundaries from raw material extraction through end-product manufacturing. For the lowest available value chain stage, Nios applies cradle-to-gate intensity factors, which capture all upstream value chain impacts (analogous to upstream [Scope 3](https://ghgprotocol.org/corporate-value-chain-scope-3-standard) in GHG accounting). For all higher-level stages, Nios applies gate-to-gate intensity factors, which include only direct, stage-specific impacts (analogous to [Scope 1](https://ghgprotocol.org/sites/default/files/ghgp/standards_supporting/Diagram%20of%20scopes%20and%20emissions%20across%20the%20value%20chain.pdf) in GHG accounting). This approach ensures that upstream impacts are fully accounted for at the most disaggregated stage while avoiding double counting across higher-level stages. Downstream stages such as product use and end-of-life are not yet included. ### Impact group coverage Nios has chosen impact groups based on their relevance to known regulations, standards, and science-based frameworks such as the Planetary Boundaries and [IPBES](https://www.unep.org/news-and-stories/story/five-drivers-nature-crisis#:~:text=Changes%20in%20land%20and%20sea,habitat%20due%20to%20global%20warming).
| Impact Group | Impact Categories | Metric | Explanation | | ----------------- | ----------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Biodiversity** | Climate change
Land use
Water consumption
Pollution | **PDF·yr (Potentially Disappeared Fraction of species over time globally)** | Risk indicator based on the fraction of species committed to global extinction over time due to human pressures. Based on spatially explicit LC-IMPACT characterization factors. Sub-categories highlight which pressures drive biodiversity loss. | | **GHG emissions** | Not further subdivided | **kg CO₂e (kilograms of carbon dioxide equivalents)** | Represented as aggregated greenhouse gases characterized into CO₂ equivalents using IPCC 100-year global warming potentials. | | **Land use** | Annual crops
Permanent crops
Pastures
Extensive forestry
Intensive forestry | **m² (square meters of land occupied)** | Group measuring the area of land used. Categories reflect land-use types with distinct ecological impacts (e.g., cropland vs. forestry). | | **Water use** | Blue water (agriculture)
Blue water (non-agriculture) | **m³ H₂Oe (cubic meters of water equivalents)** | Group capturing the absolute physical volume of freshwater withdrawn or consumed. "H₂Oe" normalizes different water flows (e.g., surface, groundwater) into a common comparable unit. Reflects demand regardless of local scarcity. | | **Water stress** | Agriculture
Non-agriculture | **m³ H₂Oe (cubic meters of water equivalents, adjusted for scarcity)** | Group adjusting water use by local scarcity and competition factors. Using 1 m³ in water-scarce regions counts for more than in water-abundant regions. Highlights where water demand is most critical for ecosystems and people. | | **Pollution** | SO₂
NOₓ
NH₃ | **kg of emissions (SO₂, NOₓ, NH₃)** | Captures emissions that typically drive terrestrial acidification. Expressed as kilograms of each pollutant released. |
## Scientific frameworks and data Nios calculations are based on the following scientific frameworks and datasets: | Scientific framework and data | Description | Comments | | ------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **GLORIA** | GLORIA (Global Resource Input-Output Assessment) is an Environmentally Extended Multi-Regional Input-Output (EE-MRIO) dataset constructed by [The University of Sydney's IELab](https://ielab.info/labs/ielab-gloria) and commercialized by [FootprintLab](https://www.footprintlab.io/).

Comprehensive technical description of the GLORIA database can be found [here](https://scp-hat.org/wp-content/uploads/2021/11/Technical-Documentation_GLORIA_20210913.pdf?utm_source=chatgpt.com).

More information about the underlying dataset used in Nios App can be found [here](https://coda.io/d/FootprintLab-Documentation-Home_dc_5ZGGkLQs/FootprintLab-Documentation_suxUkLwO#_luOotRqM). | [GLORIA is recognized as one of the most comprehensive EE-MRIO databases globally](https://www.iioa.org/conferences/31st/papers/files/5315_\[fullpaper]2025IIOAimportfactorscomparison_5315.pdf), covering \~160 countries plus 4 Rest-of-World regions and 120 economic sectors, and providing spend- and sector-based intensity factors for GHGs, land use, water use, and pollution.

It has been [selected by the United Nations (UN) to monitor progress on Sustainable Development Goals](https://unstats.un.org/sdgs/metadata/files/Metadata-08-04-01.pdf) (SDG 8 and 12). Data has also been used elsewhere by the UNEP's [Sustainable Consumption and Production](https://scp-hat.org/methods/) project, and is used by many multi-national companies.

It has been built on reputable global data sources including OECD, EDGAR, FAO, Aquastat, and others, ensuring internationally recognized quality and consistency of environmental extensions. | | **[LC-Impact](https://lc-impact.eu/methodology.html)** | LC-Impact is a life cycle impact assessment (LCIA) methodology that translates environmental pressures into biodiversity impacts (PDF score). LC-Impact was launched under the [European Union's FP7 funding program](https://www.lc-impact.eu/about.html) and is designed to align with globally recognized LCA frameworks such as the [UNEP-SETAC Life Cycle Initiative](https://www.lifecycleinitiative.org/). | Unlike many LCIA methods, [LC-Impact is spatially explicit](https://onlinelibrary.wiley.com/doi/10.1111/jiec.13018), meaning characterization factors vary by ecosystem type and geographic region, making it one of the most advanced methods for assessing biodiversity loss in LCA.

To calculate PDF metrics, we follow the biodiversity impact translation scheme used in [the Kulionis et al. (2024) research paper](https://onlinelibrary.wiley.com/doi/10.1111/jiec.13515). Note: for now, we exclude N and P emission factors as these are not provided in the GLORIA dataset. | ## Note on alignment with global standards In nature footprinting and broader supply chain impact assessment, global standards are still evolving. At this stage, Nios is not formally certified under any single standard. Our methodology is therefore designed to align closely with the shared core principles of leading frameworks and scientific approaches, and is continuously refined as standards, regulatory guidance, and scientific understanding mature * Mimics core principles of life cycle assessment (LCA) to ensure scientific grounding and transparency. * Prioritizes pragmatism by providing actionable insights for decision-makers today rather than waiting for perfect standardization. * Covers environmental impacts holistically (not only GHGs), in line with the broader framing of the [Planetary Boundaries Framework](https://www.stockholmresilience.org/research/planetary-boundaries.html) and [IPBES](https://www.unep.org/news-and-stories/story/five-drivers-nature-crisis#:~:text=Changes%20in%20land%20and%20sea,habitat%20due%20to%20global%20warming) drivers of nature loss. We monitor the development of key frameworks such as [CSRD ESRS](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32022L2464), [TNFD](https://tnfd.global/), [SBTN](https://sciencebasedtargetsnetwork.org/), [EU PEF](https://eplca.jrc.ec.europa.eu/EnvironmentalFootprint.html), [ISO](https://www.iso.org/standard/37456.html?utm_source=chatgpt.com), and the [GHG Protocol](https://ghgprotocol.org/). We aim for conceptual consistency with these standards where possible and may revisit our stance as alignment between standards matures. Our methodological choices already overlap with the recommendations of these frameworks: * Environmentally-extended input-output tables (EEMRIOs) are repeatedly referenced in the [TNFD LEAP methodology](https://tnfd.global/wp-content/uploads/2023/08/Guidance_on_the_identification_and_assessment_of_nature-related_Issues_The_TNFD_LEAP_approach_V1.1_October2023.pdf?v=1698403116) (e.g., pages 44, 84, 89, 236, 237) as a way to model value chain impacts when company-specific data is unavailable. * The [TNFD biodiversity footprinting discussion paper](https://tnfd.global/wp-content/uploads/2023/12/Discussion-paper-on_Biodiversity-footprinting-approaches-for-financial-institutions_2023.pdf?v=1701785880) explicitly refers to LC-Impact (pages 25, 42, 48). ## Limitations and future work * **Exclusions:** Current scope does not yet include various other impact groups and categories such as ecotoxicity, invasive species, ocean acidification, or marine biodiversity impacts. * **Ecosystem weighting:** Current results aggregate impacts across terrestrial, freshwater, and marine ecosystems without applying explicit weighting between ecosystem types. This reflects the methodology used in the [Kulionis et al. (2024) research paper](https://onlinelibrary.wiley.com/doi/10.1111/jiec.13515) research paper, which Nios follows for calculating PDF-based metrics. The question of how different ecosystems should be weighted is still evolving in the scientific and policy community, and we continue to monitor developments and may introduce ecosystem-specific views or alternative weighting approaches as methodologies mature. * **Averaging:** Sector-based models assume uniform technologies and sectoral-level system boundaries, which may under- or over-estimate company/product-level results. The resulting values should therefore be interpreted as directional approximations rather than precise footprints. Where available, primary data or activity/process-level LCAs can provide more granular and company-specific results. * **Inventory data:** We will expand support for mass- and quantity-based inventory data (e.g., process-level inputs, flow quantities, and detailed impact intensity factors). Current results rely primarily on spend- or sector-based intensity factors, and future updates will improve granularity and precision. * **Agent limitations:** The agent is non-deterministic and relies on probabilistic reasoning. While designed to act as a specialist in matching and analyzing data, results may vary between runs and occasional mismatches can occur. Outputs are structured approximations meant to support (not replace) expert review. # Overview (/docs/impact-measurement/overview) Follow these two steps to measure your products' environmental impacts: Create your products' bill of materials to set up the foundation for impact calculation. Calculate and analyze environmental impacts of your products. The impact measurement workflow is driven through [chat](/docs/chat/overview) in [Agent mode](/docs/chat/chat-modes#agent-mode). You can organize your impact work into [projects](/docs/projects/overview) with persistent files and instructions. See [Methodology](methodology) to understand the scientific framework, data used, and calculation methods behind impact measurement. # Project files (/docs/projects/files) Project files are documents you upload to a project. Once processed, they become part of the project's knowledge base: the agent automatically searches across all project files when answering questions, without you needing to manually reference each file. ## Upload files Drag and drop files into the project's files area, or use the file picker to select files from your device. You can upload multiple files at once. **Supported file types:** | Type | Formats | Max file size | | ------------- | ------------------------------- | ------------- | | Spreadsheets | Excel (.xlsx, .xls), .csv, .tsv | 50 MB | | Documents | PDF (.pdf) | 50 MB | | Documents | Word (.docx) | 20 MB | | Text and data | .txt, .md, .json | 100 MB | Large files are automatically truncated to fit the agent's context window. You'll see an indicator when a file has been truncated. ## File processing After upload, each file goes through a processing pipeline (extraction, chunking, and embedding) that prepares it for automatic retrieval. You can track the status of each file: * **Processing:** The file is being prepared. This may take a moment for large files. * **Ready:** The file is fully indexed and available to the agent. * **Error:** Something went wrong during processing. Try reuploading the file. ## How the agent uses project files The agent automatically retrieves relevant content from your project files when responding to messages in any project chat. This means you don't need to attach or reference files manually for the agent to use them. For more control, you can [@-mention specific files](/docs/chat/file-references) in a chat message to ensure the agent prioritizes them in its response. Files from [parent projects](/docs/projects/overview#project-hierarchy) are also available. The agent can retrieve from them automatically, and you can @-mention them in any child project chat. When the agent uses content from a project file, it includes a [citation](/docs/chat/sources-and-citations) linking back to the source, so you can verify the information. ## Preview project files Click a file in the file list to preview the extracted content. The preview shows exactly what the agent sees when retrieving from the file. For spreadsheets with multiple sheets, the preview includes tabs to navigate between sheets. ## Best practices The agent retrieves the most relevant sections from your project files for each message. The more focused and relevant your files are, the better the retrieval quality. ### Keep files focused Upload files that contain data directly relevant to the project's scope. For example, if your project focuses on a specific product line, upload the supplier reports for that product line rather than a company-wide report covering all products. This helps the agent retrieve the right content more consistently. ### Organize by project Use the [project hierarchy](/docs/projects/overview#project-hierarchy) to organize files at the right level. Upload shared reference data (company policies, standard methodologies) to a parent project so it's available across all child projects. Upload product-specific data to the relevant child project. ## Manage files * **View:** See all uploaded files and their processing status in the project's files area. Click a file to preview its content. * **Delete:** Remove a file from the project. The agent will no longer retrieve content from it. Existing citations in chat history remain visible. * **Re-upload:** If a file was updated externally, upload the new version. Delete the old version manually to avoid duplicate content. ## Project files vs. chat file attachments | | Project files | Chat file attachments | | ------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------- | | **Scope** | All chats in a project. | Only the chat the file is attach to. | | **Retrieval** | Automatically retrieved based on your message content. [@-mention](/docs/chat/file-references) specific files to prioritize them for retrieval. | By attaching the file to the chat. | | **Context** | Indexed and retrieved fresh for each message from the project files. | Available in follow-up messages. The agent prioritizes recent attachments in longer conversations. | | **Best for** | Recurring reference data organized by project. | One-off analysis in a specific chat. | # Project instructions (/docs/projects/instructions) Project instructions are persistent, user-defined guidance that the agent follows in every chat within the project. Use them to set expectations, focus the agent's analysis, or establish conventions that apply to all conversations. ## Writing instructions Open the project instructions editor from the project settings or the chat view to write or edit your instructions. You can @-mention [project files](/docs/projects/files) within your instructions to give the agent standing guidance about specific documents. For example, you might instruct the agent to always reference a particular audit report when discussing compliance. ## How instructions work Instructions are included automatically in every chat message within the project. You don't need to repeat them in each conversation. The agent follows them consistently. The agent follows instructions in both [Agent mode and Ask mode](/docs/chat/chat-modes). Instructions have high priority in the agent's context: they are preserved even in longer conversations where older messages may be condensed. Sub-projects inherit instructions from their [parent projects](/docs/projects/overview#project-hierarchy). This lets you set shared conventions at a higher level and refine them with more specific guidance in child projects. ## Example instructions ### Scope and focus ``` Focus analysis on Scope 3 emissions and biodiversity impacts. ``` ### Conventions ``` Use metric tons as the default unit for all material quantities. ``` ### File-specific guidance ``` Always reference @supplier-audit-2025.pdf when discussing compliance risks. ``` ### Comparisons ``` When comparing materials, always include cost, GHG, and biodiversity metrics. ``` ## Best practices ### Keep instructions specific Vague instructions like "be thorough" are less effective than specific directives like "always include cost, GHG, and biodiversity metrics when comparing materials." The agent treats instructions as standing rules, so write them as clear, actionable guidance. ### Use @-mentions for file-specific guidance When an instruction references a specific document, @-mention the file so the agent can retrieve its content automatically. This is more reliable than referring to files by name in plain text. ### Use project hierarchy for shared conventions Set company-wide standards (units, metrics, reporting format) in a parent project's instructions. Refine with product-specific or analysis-specific guidance in child projects. This keeps conventions consistent while allowing flexibility where needed. Instructions guide the agent at the project level. For one-off guidance that applies to a single message, include it directly in your chat message instead. # Managing projects (/docs/projects/managing-projects) ## Create a project 1. Click **New project** in the Projects sidebar. 2. Enter a project name. 3. Optionally add a description. After creation, you can start adding [files](/docs/projects/files) and [instructions](/docs/projects/instructions) to build your project's knowledge base. ## Organize with sub-projects Nest projects as sub-projects for hierarchical organization. This is useful when you have a broad topic (e.g., "Product Line Assessment") with multiple distinct workstreams underneath. To create a sub-project, select the parent project first, then create a new project within it. Child projects automatically inherit files and instructions from their parents. See [Project hierarchy](/docs/projects/overview#project-hierarchy) for details. ## Move chats into a project You can move a standalone chat into a project from the chat sidebar. Once moved, the chat gains access to the project's files and instructions. See [Managing chats](/docs/chat/managing-chats) for more on organizing conversations. ## Edit project settings Update a project's name or description at any time from the project settings. ## Delete a project Deleting a project removes the project and its associated files and instructions. Chats within the project are not deleted but become standalone chats. Deleting a project also deletes all uploaded project files and instructions. This action cannot be undone. # Overview (/docs/projects/overview) A **project** is a container that groups related chats, files, and instructions around a specific topic, product, or workflow. Projects let you build a focused knowledge base: the agent in any project chat automatically draws on project files and follows project instructions without you needing to repeat context each time. ## What a project contains * **Chats:** Conversations scoped to the project. Each chat inherits the project's files and instructions. * **Files:** Uploaded documents (spreadsheets, reports, PDFs) that serve as automatic context for all project chats. See [Project files](/docs/projects/files). * **Instructions:** Persistent guidance that shapes the agent's behavior across all project chats. See [Project instructions](/docs/projects/instructions). ## Project hierarchy You can nest projects as sub-projects to organize complex work. For example, a "Product Line Assessment" project might contain sub-projects for individual products, each with its own files and instructions. Child projects inherit context from their parents: [files](/docs/projects/files) uploaded to a parent project are available for retrieval in all child project chats, and [instructions](/docs/projects/instructions) set at a parent level apply to all child projects. This lets you share common reference data and conventions across related workstreams. ## When to use a project | Scenario | Use | | --------------------------------------------------- | -------------------------------- | | Quick one-off question | Standalone chat | | Ongoing work on a specific product or topic | Project | | Multiple analyses that share the same supplier data | Project with shared files | | Different team members working on related analyses | Project with shared instructions | Start with a standalone chat for ad-hoc questions. When a conversation evolves into sustained work, [move it into a project](/docs/projects/managing-projects) to gain access to persistent files and instructions. ## Key features Set up projects, organize them with sub-projects, and manage settings. Upload and manage files for automatic context retrieval. Define persistent instructions that guide the agent. # Data requirements (/docs/workspace/data-requirements) # Data requirements Content coming soon. # Importing files (/docs/workspace/importing-files) # Importing files Content coming soon. # Workspace overview (/docs/workspace/overview) # Workspace overview Content coming soon.