llms.txt
Compare with other sites' llms.txt files →# Agno
> Agno is an agent platform for customer-facing and internal agents. Build agents, teams and workflows with the Agno SDK, run them on AgentOS in your own infrastructure, and manage them from the Control Plane. Serve agents through your product over REST and through MCP clients.
Use this index to choose what to read. Follow the linked indexes rather than guessing page URLs. Summaries help select a source; consult the linked page for details and the documentation for version-specific APIs.
To set up Agno or build a new agent, start with [Set up Agno](https://www.agno.com/start.md).
To serve agents you already have, see [Run existing agents](https://www.agno.com/run-existing-agents.md).
For a specific use case, choose its Build guide below. Product pages explain the components.
## Build
- [Set up Agno and build your first agent](https://www.agno.com/start.md): Type: Build. Choose a deployment target, run the shared setup prompt, and verify your first agent.
- [Build a docs agent](https://www.agno.com/build/docs-agent.md): Type: Build. Build an assistant grounded in your documentation and serve it over MCP.
- [Serve agents to your end users](https://www.agno.com/build/end-user-management.md): Type: Build. Plan identity, per-user data isolation, roles and authorization for a customer-facing agent.
- [Put an agent in your product](https://www.agno.com/build/in-product-agent.md): Type: Build. Connect an agent to your application with authenticated sessions, streaming and controlled actions.
- [Run agents you already have](https://www.agno.com/run-existing-agents.md): Type: Build. Evaluate serving existing agents on AgentOS before deciding whether to port their implementation.
## Learn
- [Agno SDK](https://www.agno.com/products/sdk.md): Type: Conceptual. Python primitives for agents, teams and workflows, with tools, knowledge and memory.
- [AgentOS runtime](https://www.agno.com/products/runtime.md): Type: Conceptual. Serve agents as an API in your own infrastructure and operate their sessions and runs.
- [Control Plane](https://www.agno.com/products/control-plane.md): Type: Conceptual. Manage agents, inspect sessions and traces, and review approvals and evaluations.
## Knowledge base
- [Agno vs. Vercel AI SDK: choosing a stack for your AI product](https://www.agno.com/kb/guide/agno-vs-vercel-ai-sdk.md): Type: Comparison. Agno is a Python framework with a runtime you self-host. The Vercel AI SDK is a set of TypeScript libraries. Explore how they compare on memory, durable execution, access controls, and harnesses.
- [Why can’t the Control Plane connect to my AgentOS?](https://www.agno.com/kb/guide/control-plane-cant-connect-to-agentos.md): Type: Reference. The Control Plane connects from your browser, not Agno's servers. Here are the six reasons a connection fails, how to tell which one you have, and how to fix each.
- [How to price an AI feature in your SaaS product](https://www.agno.com/kb/guide/price-an-ai-feature-in-your-saas-product.md): Type: Reference. Four pricing models for AI features, two worked examples with the cost arithmetic, why per-token pricing fails, and how to price with headroom.
- [How to fix a rejected Claude connector or ChatGPT plugin](https://www.agno.com/kb/guide/when-a-directory-rejects-your-submission.md): Type: Reference. Troubleshoot rejected MCP connectors and plugins for Claude and ChatGPT. Fix connection, policy, tool, and annotation issues, then resubmit.
- [Agentic workflows vs. AI agents: when to use each](https://www.agno.com/kb/guide/agentic-workflows-vs-ai-agents.md): Type: Reference. A workflow follows steps you define, an agent decides its own steps. How they differ, when to use each, and how to build both in Agno.
- [Agno tracing vs. LangSmith](https://www.agno.com/kb/guide/agno-tracing-vs-langsmith.md): Type: Comparison. Agno writes OpenTelemetry spans to your own database with no per-trace fee. LangSmith is a hosted platform with deeper evals, billed per trace and capped at 180 days.
- [Agno durable execution vs. Pydantic AI Temporal](https://www.agno.com/kb/guide/durable-execution-agno-vs-pydantic-ai.md): Type: Comparison. Pydantic AI hands a run to an external engine like Temporal or DBOS. Agno checkpoints to the database you already run and resumes in place. What each guarantees, and when you need an engine.
- [How to choose a model for each agent task in Agno](https://www.agno.com/kb/guide/choose-a-model-for-each-agent-task.md): Type: Reference. Pick a model per agent: reasoning depth, tool calling, context, latency, cost, and where the data can go, with fallbacks for when a provider fails.
- [How to connect an Agno agent to Slack](https://www.agno.com/kb/guide/connect-an-agno-agent-to-slack.md): Type: Reference. Add the Slack interface to AgentOS so your agent answers DMs and @mentions, including inside threads, as each user. Setup takes a Slack app manifest, two environment variables, and a few lines of code.
- [How to publish your agent to the ChatGPT plugin directory](https://www.agno.com/kb/guide/publish-your-agent-to-the-chatgpt-directory.md): Type: Reference. A ChatGPT plugin wraps the MCP server you already run. Here's what OpenAI checks: OAuth, tool annotations, test cases, and a verified publisher.
- [How to ship your agent as a Claude connector](https://www.agno.com/kb/guide/ship-your-agent-as-a-claude-connector.md): Type: Reference. A Claude connector is a remote MCP server Claude can call. Add OAuth, tool metadata, and testing to your MCP server, then list it in the directory.
## Choosing and comparing
- [Compare agent frameworks](https://www.agno.com/compare.md): Type: Comparison. Find a comparison, including strengths, tradeoffs and sources.
- [Agno vs LangGraph](https://www.agno.com/compare/agno-vs-langgraph.md): Type: Comparison. Compare capabilities and when to choose each; source review dated 2026-08-10.
- [Agno vs CrewAI](https://www.agno.com/compare/agno-vs-crewai.md): Type: Comparison. Compare capabilities and when to choose each; source review dated 2026-08-10.
- [Agno vs OpenAI Agents SDK](https://www.agno.com/compare/agno-vs-openai-agents-sdk.md): Type: Comparison. Compare capabilities and when to choose each; source review dated 2026-08-10.
## Pricing and enterprise
- [Pricing](https://www.agno.com/pricing.md): Type: Reference. Free, Pro and Enterprise plans, included capabilities and add-ons.
- [Enterprise](https://www.agno.com/enterprise.md): Type: Reference. Enterprise deployment, governance, support and contact information.
## Evidence
- [Benchmarks](https://www.agno.com/benchmarks.md): Type: Reference. Agent instantiation time and memory, with hardware, versions, measurement date and reproduction harness.
- [Customers](https://www.agno.com/customers.md): Type: Reference. Production deployments, attributed customer accounts and linked case studies.
- [Engineering blog](https://www.agno.com/blog.md): Type: Reference. Full index of engineering guides and case studies, with publication dates.
## Documentation and reference
- [Canonical names and naming history](https://www.agno.com/taxonomy.json): Type: Reference. Canonical product names, aliases, historical names such as Phidata, and sourced deprecations with replacements.
- [Documentation index](https://docs.agno.com/llms.txt): Type: Reference. SDK and AgentOS guides, examples and API reference; consult for implementation details.
- [Docs MCP server](https://docs.agno.com/mcp): Type: Reference. Query the Agno documentation from a coding agent that supports MCP.
- [Source code and cookbook](https://github.com/agno-agi/agno): Type: Reference. Agno source, working examples and release history.
- [Changelog](https://www.agno.com/changelog.md): Type: Changelog. Dated feature and fix entries, with release versions where available.
## Optional
- [Full page index](https://www.agno.com/sitemap.md): Type: Reference. Every listed page, build guide and published article, with type and summary.
- [About](https://www.agno.com/about.md): Type: Conceptual. The team behind Agno.
- [Community](https://www.agno.com/community.md): Type: Reference. Community support and places to participate.
- [Full site content for LLMs](https://www.agno.com/llms-full.txt): Type: Reference. An expanded content export; use the page index for targeted retrieval.