First-party ecosystem

Bring the providers you choose.

Models, search, speech, channels, tools, agent runtimes, and complete interfaces are apps. Install what you need. Replace it when your needs change.

Open the app repository
38Manifest-backed apps
9App categories
1Permission-aware contract
0Required cloud vendors

38 apps

Models

Alibaba Model Studio

Qwen chat, embedding, image generation, video generation, and more through DashScope.

  • Qwen
  • Embedding
  • Image
  • Video
Models

Anthropic-Compatible

Connect any endpoint that implements the Anthropic Messages API.

  • Chat
  • Compatible
Models

Anthropic

Claude chat models through the Anthropic Messages API.

  • Claude
  • Chat
Models

Amazon Bedrock

Chat, embedding, image, video, and speech models using your AWS credential chain.

  • Chat
  • Embedding
  • Media
  • Speech
Search

Brave Search

Broad web search with ranked links, snippets, and a recency filter.

  • Web
  • Recency
Agents

Claude Code

Run Claude Code over ACP with every tool call routed through the host approval gate.

  • ACP
  • Coding
  • Approvals
Agents

OpenAI Codex

Run the Codex CLI over ACP while PersonalClaw enforces host tool approvals.

  • ACP
  • Coding
  • Approvals
Models

DeepSeek

DeepSeek chat and reasoner models through its OpenAI-compatible API.

  • Chat
  • Reasoning
Speech

Diarization (ONNX)

Install-and-go local speaker diarization using sherpa-onnx.

  • Local
  • Diarization
Speech

Diarization (pyannote)

High-accuracy speaker diarization using the pyannote.audio pretrained pipeline.

  • Diarization
  • Local
Search

DuckDuckGo

Keyless, zero-config web search with links and snippets.

  • Web
  • Keyless
Search

Exa

Semantic web search with relevant highlights and extracted content.

  • Semantic
  • Extract
Models

FAL

Hosted image and video generation across FAL's model catalog.

  • Image
  • Video
Speech

Faster Whisper

Local speech-to-text with CTranslate2-optimized Whisper models.

  • STT
  • Local
Models

Google Gemini

Gemini chat, embedding, image, video, and text-to-speech models.

  • Chat
  • Embedding
  • Media
Models

Groq

Low-latency chat inference using Groq's OpenAI-compatible API.

  • Chat
  • Fast
Products

Growth Tracker

Turn real work into evidenced growth artifacts and a shareable accomplishment document.

  • Career
  • Projects
  • Tasks
Agents

Kiro CLI

Run the Kiro CLI agent over ACP when the local binary is available.

  • ACP
  • Coding
Tools

MCP Tool Servers

Connect Model Context Protocol servers over stdio or SSE transport.

  • MCP
  • stdio
  • SSE
Models

Meta Muse Spark

Meta Muse Spark chat models through the Meta AI API.

  • Chat
  • Compatible
Products

Minutes

Combine recordings, notes, transcripts, decisions, and action items on one meeting timeline.

  • Meetings
  • Transcription
  • Tasks
Models

Mistral AI

Mistral models through its OpenAI-compatible endpoint.

  • Chat
  • Compatible
Models

Ollama

Use local or remote Ollama models for chat and embedding.

  • Chat
  • Embedding
  • Local
Models

OpenAI-Compatible

Connect a self-hosted, proxied, or unlisted Chat Completions endpoint.

  • Chat
  • Compatible
Models

OpenAI

OpenAI chat, vision, code-tool, and embedding models.

  • Chat
  • Vision
  • Embedding
Tools

OpenAI Tool Servers

Connect OpenAI-compatible tool servers that expose tools over REST.

  • Tools
  • REST
Search

Perplexity Sonar

Answer-first live web search with synthesized responses and cited sources.

  • Answers
  • Citations
Speech

Piper TTS

Local neural text-to-speech with downloadable Piper voices.

  • TTS
  • Local
Search

SearXNG

Private, keyless meta-search through your own SearXNG instance.

  • Self-hosted
  • Keyless
Models

Sentence Transformers

Run text embedding models in-process on your own machine.

  • Embedding
  • Local
Skills

Skills.sh Marketplace

Search, preview, install, and manage community skills from skills.sh.

  • Skills
  • Marketplace
Channels

Slack Channel

Monitor Slack channels, respond to mentions, and participate in threads.

  • Messaging
  • Threads
Search

Tavily

Agent-oriented search with synthesized answers, scored results, and page extraction.

  • Answers
  • Extract
Models

Together AI

Serverless model inference through Together AI's compatible API.

  • Chat
  • Compatible
Models

vLLM

Connect a local vLLM server through its OpenAI-compatible endpoint.

  • Chat
  • Local
Tools

Native Tools (Web)

Give agents guarded web search, fetch, and structured extraction primitives.

  • Search
  • Fetch
  • Extract
Actions

HTTP Webhook Hooks

Send templated agent lifecycle events to an HTTP endpoint.

  • Webhook
  • Automation
Search

Wikipedia Search

Keyless encyclopedic search with ranked articles and introductory extracts.

  • Reference
  • Keyless

The app boundary

Installation is a security decision.

Apps do not become active the moment code arrives. PersonalClaw separates acquisition, inspection, consent, and enablement.

  1. 01

    Quarantine

    The package lands outside the active runtime while its manifest and contents are inspected.

  2. 02

    Scan

    Supply-chain checks assign a verdict. Dangerous findings cannot be overridden.

  3. 03

    Consent

    The owner sees requested capabilities before deciding whether to continue.

  4. 04

    Enable

    The app receives scoped access through the gateway and begins in a bounded state.

For builders

The ecosystem is an open surface.

Provider apps can be small Python adapters. Full apps can ship a backend, mounted React interface, storage, cron work, and event access behind declared permissions.

app.json{ "name": "your-app", "provider": { "type": "tool" }, "permissions": { "network": false } }
Read the app creation guide