# AirPrompter > Turn the AI instructions that work into reusable prompts and workflows you can apply to real work every day. AirPrompter keeps the AI instructions that work as saved prompts and repeatable workflows, and runs them from any AI tool through MCP, SDKs, the CLI, and a browser extension. Human-readable documentation lives at https://airprompter.com/docs; every page below is also served as clean Markdown at the linked `.md` URL. The whole documentation set in one file: https://airprompter.com/docs/llms-full.txt ## Start here - [AirPrompter documentation](https://airprompter.com/docs/index.md): Turn the AI instructions that work into reusable prompts and workflows you can apply to real work every day. - [Create your first reusable workflow](https://airprompter.com/docs/quickstart.md): Build a two-step meeting follow-up system, test its accepted result, and know exactly where to repair a failure. ## Build your system - [Build a reusable prompt library](https://airprompter.com/docs/prompts.md): Write focused prompts, provide the values that change, test realistic inputs, and keep your Library useful as your work evolves. - [Design and run workflows](https://airprompter.com/docs/workflows.md): Turn a recurring job into distinct responsibilities with explicit handoffs, review gates, and one obvious repair point. ## Run and connect - [Use AirPrompter in your browser](https://airprompter.com/docs/browser-extension.md): Find saved workflows beside a supported AI chat, fill reusable variables, choose manual or Run All execution, and stop bad context before it compounds. (availability varies by plan) - [Connect an AI host to AirPrompter](https://airprompter.com/docs/connect.md): Choose the right connector, authorize it safely, prove workflow discovery works, and understand what the host—not AirPrompter—executes. (availability varies by plan) - [Connect through MCP directly](https://airprompter.com/docs/mcp.md): Understand AirPrompter's remote MCP boundary, choose OAuth or an approved headless credential, configure a qualified host, and verify the connection safely. (availability varies by plan) ## Develop - [Choose the AirPrompter API surface](https://airprompter.com/docs/api.md): Decide between the customer REST API, MCP, Connect, and the browser extension, then make a verified request without assuming unsupported endpoints. (availability varies by plan) - [Create and protect API keys](https://airprompter.com/docs/api-keys.md): Choose the correct target, content environment, and scope; store the secret once; test least privilege; and rotate or revoke it safely. ## Integrate - [Integrate Team Agents with your own systems](https://airprompter.com/docs/integrations.md): Understand what an integration is, which public library does what, the four ways to run approved prompts from your own runtime, and how to pick one. - [Run approved prompts from Node and TypeScript](https://airprompter.com/docs/integrations/typescript.md): Install the Agent SDK, start a resident runtime, render a prompt, wrap your OpenAI or Anthropic client so every call is measured and carries the release's inference settings, and expose health. - [Run approved prompts from Python](https://airprompter.com/docs/integrations/python.md): Install the Python Agent SDK, start a resident runtime, render a prompt, wrap your openai or anthropic client, attribute calls explicitly, and instrument LiteLLM with the release's inference settings. - [Distribute releases through your own database](https://airprompter.com/docs/integrations/fleet.md): Run one puller that holds the Agent key and writes sealed release rows, boot every runtime from the newest row with only a distribution key, and understand the pointer-first pull, restart, and rollback behaviour. - [Understand and route telemetry](https://airprompter.com/docs/integrations/telemetry.md): Know exactly what a spool row contains and what it never can, choose between the in-process uploader, the host daemon, and your OpenTelemetry collector, keep the disk budget honest, and run inside a VPC with no egress. - [Call a hosted environment with the OpenAI or Anthropic SDK](https://airprompter.com/docs/integrations/hosted.md): Point the provider SDK you already use at a hosted Team Agent environment, name the prompt as the model, and know what is accepted, ignored, and refused. - [Operate hosts with the CLI and the daemon](https://airprompter.com/docs/integrations/cli.md): Pull, verify, apply, and inspect releases from a shell, run one sync loop per host with airprompterd under systemd, launchd, Docker, or Kubernetes, and carry releases and telemetry across an air gap. ## Your account - [Set your account preferences](https://airprompter.com/docs/account/preferences.md): Update your profile and authoring style, choose the default Personal or Team surface for this browser, and control guided tours without changing access or billing. ## Optional - [Pricing](https://airprompter.com/pricing): Plans and what each includes - [Blog](https://airprompter.com/blog): Articles on prompt systems and AI productivity - [Brand](https://airprompter.com/brand): Logo and brand guidelines