# AgentParley

> AgentParley runs crews of expert AI agents that talk, delegate and run long missions, with a marketplace of plugins. These docs explain how to build plugins and companions and use the API.

All of these docs in one file: https://agentparley.ai/llms-full.txt

## Start

- [LLM start here](https://agentparley.ai/docs/llm-start-here.md): Plain-text versions of these docs for AI assistants and agents, and how to point yours at them.
- [Quickstart — your first plugin](https://agentparley.ai/docs/plugin-quickstart.md): Write a plugin, publish it, install it, and watch it change an agent's reply.
- [Plugin SDK](https://agentparley.ai/docs/plugin-sdk.md): What a plugin is, what it can change, and how the pieces fit.

## Concepts

- [Sandbox model](https://agentparley.ai/docs/sandbox-model.md): What your plugin can and cannot reach, and the memory, time and network limits it runs under.
- [Permissions](https://agentparley.ai/docs/plugin-permissions.md): Permissions come from the code you wrote. How they're derived, what the installer sees, and how updates are gated.
- [Publishing](https://agentparley.ai/docs/publishing.md): Drafts, immutable versions, private-first listing, yanking, and how installs follow updates.

## Reference

- [Package format](https://agentparley.ai/docs/plugin-package-format.md): The folder layout, readme.md, the entry file, TypeScript, and the size limits a package must meet.
- [Manifest reference](https://agentparley.ai/docs/plugin-manifest.md): Every key of globalThis.manifest, what it does, and how it is validated.
- [Host API reference](https://agentparley.ai/docs/plugin-host-api.md): Every member of the host object, its arguments, its result, and when it is present.
- [API reference](https://agentparley.ai/docs/api-reference.md): The REST endpoints for publishing and installing plugins, and the webhook address plugin channels receive on.

## Guides

- [Tools](https://agentparley.ai/docs/plugin-tools.md): Give your agents a new action the model can call: a description, a JSON schema and a handler, running in the plugin sandbox.
- [Middleware hooks](https://agentparley.ai/docs/plugin-middleware.md): The 17 points where a plugin can rewrite or veto what an agent is doing, as it happens.
- [Lifecycle events](https://agentparley.ai/docs/plugin-lifecycle-events.md): Run code when your plugin is installed, updated or removed, and when a chat starts, is cleared or deleted.
- [Commands and skills](https://agentparley.ai/docs/plugin-commands-and-skills.md): Ship slash commands and on-demand instructions as plain markdown files — no code.
- [Channels](https://agentparley.ai/docs/plugin-channels.md): Connect agents to a chat app — people message the agent there, and it answers there.
- [Live connections](https://agentparley.ai/docs/plugin-live-connections.md): Socket-only platforms without running a server — AgentParley holds the connection and wakes your plugin only when something matters.
- [Model providers](https://agentparley.ai/docs/plugin-model-providers.md): Make your vendor's chat or decision models selectable on AgentParley, like OpenAI or Anthropic.
- [Web search and web fetch providers](https://agentparley.ai/docs/plugin-web-providers.md): Replace the engine behind the built-in web_search and web_fetch tools.
- [Compaction provider](https://agentparley.ai/docs/plugin-compaction.md): Replace how long conversations are summarized so long missions keep the thread.
- [Agent crews](https://agentparley.ai/docs/plugin-agent-crews.md): Ship a team, not a tool — create and heal agents, shape how they delegate, and govern what they say to each other.
- [Create a 3D companion](https://agentparley.ai/docs/create-3d-companion.md): For 3D artists. Put your character on screen in a few minutes, then share her with other people. No code needed.
- [Create a 2D companion](https://agentparley.ai/docs/create-2d-companion.md): For Live2D artists. Put your rigged character on screen in a few minutes, then share her with other people. No code needed.
- [Companion packs](https://agentparley.ai/docs/plugin-companion-packs.md): Ship 3D and 2.5D (Live2D) characters with their own motions, expressions, thinking sounds and behaviour. Data only, no code.
- [Voice and speech-to-text providers](https://agentparley.ai/docs/plugin-voice-and-speech.md): Give companions a new voice vendor, or a new way to hear the user.

## Examples

- [Example: Caveman](https://agentparley.ai/docs/example-caveman.md): The smallest real plugin — one middleware hook that changes how every agent writes.
- [Example: Brave Search](https://agentparley.ai/docs/example-brave-search.md): A bring-your-own-key web search provider — settings, secrets, egress and host.fetch.
- [Example: Pi Compactor](https://agentparley.ai/docs/example-pi-compactor.md): A structured, iterative context summary written with the agent's own model — no key, no permissions.
- [Example: Discord](https://agentparley.ai/docs/example-discord.md): A full chat channel over a held socket — resumable Gateway, filters, heartbeats, idempotent sends.
- [Example: Oh My OpenAgents](https://agentparley.ai/docs/example-oh-my-openagents.md): A seven-agent crew in one package — provisioned, self-healing, model-matched, with doctrine, commands and skills.
- [Example: Sample Companion Pack](https://agentparley.ai/docs/example-sample-companion-pack.md): A data-only pack — an avatar look with its own motions, and two ready-made characters.
- [Example: OpenAI Voice](https://agentparley.ai/docs/example-openai-voice.md): A voice and speech-to-text provider — binary audio in and out through host.fetch, typed vendor failures.

## Running crews

- [Agent packs](https://agentparley.ai/docs/agent-packs.md): Install a prebuilt crew — agents, doctrine, commands and skills — in one click.

## Tunnel

- [The tunnel](https://agentparley.ai/docs/tunnel.md): A small daemon on your own computer that lets your agents run commands there and use its local models, with no open port and no firewall change.
- [SSH commands](https://agentparley.ai/docs/tunnel-ssh.md): How an agent runs commands and edits files on your computer, what each command runs as, and the layers that decide what it may do.
- [Local models](https://agentparley.ai/docs/tunnel-local-models.md): Use Ollama, LM Studio, vLLM or llama.cpp on your own computer as a model provider for your agents, through the tunnel.
- [Codex and Claude Code](https://agentparley.ai/docs/tunnel-codex-claude.md): Run your agents on your own ChatGPT or Claude plan through the Codex CLI or Claude Code on a connected computer, with only AgentParley's tools.