--- title: "Document Processing for AI Agents | Extend" description: "Parse, extract, classify, split, and edit documents from Cursor, Claude Code, Codex, or any MCP client. MCP, CLI, SDKs, REST." canonical: https://www.extend.ai/agents --- # Document Processing for AI Agents | Extend > For agents: the docs index is [https://docs.extend.ai/llms.txt](https://docs.extend.ai/llms.txt) and the single-file platform context is [https://docs.extend.ai/agents.md](https://docs.extend.ai/agents.md). Agents ## Document processing for agents Parse, extract, classify, split, and edit documents from Cursor, Claude Code, Codex, or any MCP client. MCP, CLI, SDKs, REST. [Try for free](https://dashboard.extend.ai/) · [Agent quickstart](https://docs.extend.ai/agent-quickstart) ### MCP ```text https://mcp.extend.ai/mcp ``` Add to any MCP client. OAuth sign-in, nothing to install. ### CLI ```bash curl -fsSL https://extend.ai/install.sh | sh && extend setup ``` Installs the extend command and the extend-cli skill. ### Skill ```bash npx skills add extend-hq/extend-agent-plugin --skill extend-api ``` Teaches a coding agent the API, schema rules, and SDKs. ## Works where your agent works One hosted MCP server, seven documented clients. Pick yours for the exact config and first prompts to try. - [Cursor](https://www.extend.ai/agents/cursor) — Parse and extract documents without leaving your editor. - [Claude Code](https://www.extend.ai/agents/claude-code) — Drive every Extend processor from your terminal session. - [Claude](https://www.extend.ai/agents/claude) — Parse, extract, and fill documents from a Claude conversation. - [Codex](https://www.extend.ai/agents/codex) — Give Codex every Extend processor as a tool. - [ChatGPT](https://www.extend.ai/agents/chatgpt) — Process documents from ChatGPT with a custom connector. - [VS Code](https://www.extend.ai/agents/vscode) — Use Extend from GitHub Copilot agent mode in VS Code. - [Windsurf](https://www.extend.ai/agents/windsurf) — Parse and extract documents from Cascade in Windsurf. - [Any MCP client](https://www.extend.ai/agents/mcp) — Any client that speaks MCP over Streamable HTTP can connect. ## Pick how your agent will use Extend The three paths are independent. Set up only what your agent will use, and add another later if its role grows. ### [MCP server](https://www.extend.ai/agents/mcp) - **If your agent** If your agent calls tools from chat (Cursor, Claude, ChatGPT, and other MCP clients) - **Give it** The hosted MCP server. OAuth sign-in, nothing to install. ### [CLI and extend-cli skill](https://www.extend.ai/agents/cli) - **If your agent** If your agent runs one-off and shell-driven document jobs - **Give it** The extend CLI and the extend-cli skill generated from its command tree. ### [SDK and extend-api skill](https://www.extend.ai/agents/sdk) - **If your agent** If your agent writes durable integration code with an SDK - **Give it** The extend-api context file plus the Python, TypeScript, Java, or Go SDK. ## What to ask your agent Talk in terms of outcomes and let the agent pick the tools. Copy any prompt into your MCP client. - "Parse this PDF and summarize the key terms." - "Extract the vendor, invoice number, line items, and total from these invoices into a table." - "Classify each document in this folder as an MSA, SOW, or NDA." - "Split this scanned bundle into separate documents by page range." - "Fill out this W-9 with our company details." - "Create an extractor for utility bills, test it on these three samples, and publish it." - "Build an evaluation set from these labeled invoices and measure the extractor's accuracy." - "Set up a webhook so my app is notified when runs finish." - "List yesterday's failed extraction runs and re-run them." ## What leading AI teams say about Extend Customers: > Extend outperformed every solution we tested—other vendors, open source, and even foundation models. It now powers key document workflows across 30,000 customers, helping us build the most intelligent and modern financial platform out there. > > — Pedro Franceschi, CEO, Brex [Read more](https://www.extend.ai/resources/how-brex-reached-99-accuracy-across-millions-of-financial-documents) > Extend sets the bar for what all vendors should be - a best in class product and a world class team. We couldn't be happier and are deeply grateful for our relationship with them. > > — Yan Lhert, Staff Engineer, Opendoor > We tested AWS, foundation models, and newer vendors, and evaluated options on latency, accuracy, and the developer experience. Extend was the only one that could hit our performance bar, with the tooling we needed to scale. > > — Paul-Arthur Asselin, Senior Software Engineer, Mercury [Read more](https://www.extend.ai/resources/mercury-case-study) [See all customer stories >](https://www.extend.ai/customers) ## Frequently asked questions ### Do I need an API key to use the MCP server? No. The hosted MCP server uses OAuth: your client opens a browser sign-in and you choose which workspaces it may use and whether it acts in Test or Production. API keys are for the CLI in headless environments and for SDK code. ### Should my coding agent use MCP, the CLI, or an SDK? MCP is the best fit for chat and tool-calling clients. For agents writing durable integration code, use an SDK with the extend-api skill; for one-off shell-driven jobs, the CLI with the extend-cli skill. The paths are independent, so set up only what your agent will use. ### Can Claude or ChatGPT process files on my computer? Chat clients don't expose local files to remote servers. When you ask for a file that's only on your machine, the agent requests an upload link: open it, sign in to the Extend dashboard, drop your files, and click Done. Links expire after 15 minutes and only members of the target workspace can use them. Documents at a public URL need no upload. ### How do I keep an agent out of production? Every operation targets a specific workspace and environment, and the agent must state both on every call. On the OAuth consent screen, grant the client the Test Environment only. Resources are separate between Test and Production, so an extractor created in Test never exists in Production. ### How do I limit which tools my agent sees? Append a comma-separated list of group or tool names in the tools query parameter, for example https://mcp.extend.ai/mcp?tools=extract,files. The groups are extract, classify, parse, split, workflows, edit, files, webhooks, and evaluations. Access and documentation tools stay available. ### Can my agent read the docs directly? Yes. https://docs.extend.ai/llms.txt indexes every page, and appending .md to any docs URL returns clean Markdown. Every page on extend.ai has a Markdown twin too. Agents on the MCP path get the same through the search_documentation and get_documentation tools. ### Is there a free tier for agent projects? Yes. Pay As You Go includes 10,000 free credits to get started, then bills per credit with no monthly minimum. See pricing for the full rate table.