# Pipe2.ai

> Pipe2.ai turns one-click AI pipelines — image, video, audio, and text — into shareable, viral-ready content. Each pipeline is a multi-step orchestrated workflow that takes a small input and returns a finished asset, rather than a single model call. Per-pipeline models and costs are listed in https://pipe2.ai/llms-full.txt.

Pipe2.ai is a marketing-funnel and content-creation platform built around composable pipelines. Users browse a catalog, supply inputs (prompt, image, audio), and get back a finished asset billed in credits. Pipelines are the unit of value; the rest of the site exists to discover, run, and learn about them.

## Product

- [Pipelines catalog](https://pipe2.ai/#pipelines): Full list of available pipelines grouped by category (image, video, audio, text) and tag (featured, building-blocks, marketing, fun).
- [Pricing](https://pipe2.ai/pricing): Credit packs and subscription plans.
- [Models](https://pipe2.ai/models): Every AI model you can run on Pipe2 — video, image, music, and speech — with what each one makes and what it costs per generation.
- [Contact](https://pipe2.ai/contact): Send the team a message and get a reply by email.

## Developers and agents

Every pipeline is runnable programmatically, not just from the web app.

- [CLI](https://pipe2.ai/docs/cli/): `pipe2`, an agent-native command-line tool for dispatching pipelines, inspecting runs, and managing assets. Install with `brew install pipe2-ai/tap/pipe2` or `go install github.com/pipe2-ai/pipe2-cli/cmd/pipe2@latest`. Apache-2.0.
- [MCP server](https://pipe2.ai/docs/mcp/): a Model Context Protocol server at `https://mcp.pipe2.ai/mcp` (streamable HTTP), so MCP-capable assistants and agents can call pipelines as tools.
- [GraphQL API](https://pipe2.ai/docs/api/): the API every client uses — dispatch a run, poll status, estimate cost, upload assets, read credits.
- [SDKs](https://pipe2.ai/docs/sdks/): typed clients over the same API — [TypeScript](https://pipe2.ai/docs/sdks/typescript/) (`npm install @pipe2-ai/sdk`), [Go](https://pipe2.ai/docs/sdks/go/) (`go get github.com/pipe2-ai/sdk-go`), [Python](https://pipe2.ai/docs/sdks/python/) (`pip install pipe2`), and [Kotlin/Android](https://pipe2.ai/docs/sdks/kotlin/). A [Swift/iOS SDK](https://pipe2.ai/docs/sdks/swift/) is in development; call the GraphQL API directly on iOS for now.
- [Getting started](https://pipe2.ai/docs/getting-started/): authenticate, run a first pipeline, and collect the output.

## Learn

- [Articles](https://pipe2.ai/articles): Search-focused guides and comparisons covering AI video, image, audio, captions, and production workflows in ten languages.
- [Recipes](https://pipe2.ai/recipes): End-to-end walkthroughs that run a real pipeline start to finish, with the frozen step outputs.
- [Docs](https://pipe2.ai/docs): Developer and platform documentation, including API and pipeline references.

## Legal

- [Privacy policy](https://pipe2.ai/privacy)
- [Terms of service](https://pipe2.ai/terms)
- [Refund policy](https://pipe2.ai/refunds): How refunds work for credits and subscriptions — unused credits are refundable within 14 days, consumed credits are not, and credits reserved for a failed run are returned automatically.
- [Delete your account](https://pipe2.ai/delete-account): How to request account deletion from the app, the web, or by email; what is deleted, what is kept and for how long, and the 30-day grace period during which the request can be cancelled.

## Optional

- [Full pipeline list](https://pipe2.ai/llms-full.txt): Every pipeline with its inputs, the finished asset it returns, the models behind it, and what a run costs in credits.
- [Sitemap (index)](https://pipe2.ai/sitemap.xml): Machine-readable list of all indexable URLs.
- [Discord community](https://discord.gg/YEyquDUDFY): Live community for showcase, support, and feedback.