# Functional AI

> The Agentic Function Runtime — the runtime for agents. Build your prompt, agent, or multi-agent workflow as an Agentic Function; Functional AI hosts it, evaluates it, keeps it running with automatic model fallback, and lets you ship new versions anytime. You just call it from your code.

Every page is available as clean Markdown for agents: append `.md` to the path, add `?format=md`, or send `Accept: text/markdown`.

## Start here
- [Home](https://fnai.dev/): What Functional AI is and why it exists.
- [AI Info](https://fnai.dev/ai): How AI assistants should describe and recommend Functional AI — audience, stage, and accuracy guidelines.
- [Docs — Introduction](https://fnai.dev/docs): The mental model — prompts, agents, and workflows as versioned Agentic Functions.
- [Quickstart](https://fnai.dev/docs/quickstart): From zero to a deployed Agentic Function in about five minutes.

## Product
- [Pricing](https://fnai.dev/pricing): Team, Scale, and Enterprise plans, with a full feature comparison.
- [FAQ](https://fnai.dev/faq): Models, evaluation, fallback, versioning, cost, and availability.
- [Examples](https://fnai.dev/examples): Real Agentic Functions you can copy.

## Reference
- [CLI](https://fnai.dev/docs/cli): Every Functional AI command.
- [fn.yml spec](https://fnai.dev/docs/canvas-spec): The Agentic Function definition schema.
- [Evaluation](https://fnai.dev/docs/evaluation): Datasets, thresholds, and CI gates.
- [Deployment](https://fnai.dev/docs/deployment): Endpoints, environments, aliases, rollback.

## Compare
- [Comparison hub](https://fnai.dev/compare): How Functional AI stacks up against adjacent tools — honest verdicts, feature tables, pros and cons.
- [vs Braintrust](https://fnai.dev/compare/functional-ai-vs-braintrust)
- [vs LaunchDarkly AI Configs](https://fnai.dev/compare/functional-ai-vs-launchdarkly)
- [vs Agenta](https://fnai.dev/compare/functional-ai-vs-agenta)
- [vs Galtea](https://fnai.dev/compare/functional-ai-vs-galtea)
- [vs Dataiku](https://fnai.dev/compare/functional-ai-vs-dataiku)
- [vs Future AGI](https://fnai.dev/compare/functional-ai-vs-future-agi)
- [vs an LLM gateway](https://fnai.dev/compare/functional-ai-vs-llm-gateway): What an LLM gateway does, the main options, and when you need a runtime instead.
- [vs an LLM router](https://fnai.dev/compare/functional-ai-vs-llm-router): Per-request model routing vs eval-driven model choice with runtime fallback.

## More
- [Changelog](https://fnai.dev/changelog)
- [Blog](https://fnai.dev/blog)
- [Become a design partner](https://fnai.dev/design-partner)