AI Decision System
Replayable, offline-verifiable records for AI-assisted go/no-go decisions: experiment promotions, typed model answers and human-approved actions.
Your mobile companion for OpenClaw and Hermes. See your agents, follow conversations, and take control from iPhone, iPad, or Android.
Developer toolsTypeScript · SwiftGitHub @p697OnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Run /nomorevibe verify in your project folder to verify ownership and manage the introduction yourself.
Verify in one minuteAsk about other verification options
Mobile Client for OpenClaw, Hermes, and More
Taken from the repository README
Loading the full README. The saved excerpt is shown below.
Your agents. Always with you.
Clawket is an open-source iOS and Android companion for OpenClaw and Hermes Agent. Chat with your agents, follow their tool calls, switch between sessions, and manage their work from your phone.
App Store ↗
Google Play ↗
The screenshots show the new 3.0 interface. Store availability may differ while the release is under review.Read the full README ↗Your agents. Always with you.
Clawket is an open-source iOS and Android companion for OpenClaw and Hermes Agent. Chat with your agents, follow their tool calls, switch between sessions, and manage their work from your phone.
The screenshots show the new 3.0 interface. Store availability may differ while the release is under review.
Clawket connects to agents you run; you need an existing supported agent or model server. Available tools and management features depend on that backend.
Install the mobile app using either store link above. On the computer running your agent, install the Bridge CLI (Node.js 20.3+):
npm install -g @p697/clawket
clawket pair
Scan the generated QR code in Clawket. The default command detects installed OpenClaw and Hermes backends and produces a labeled pairing result for each. Relay is the default; Hermes pairing also attempts to start its Clawket-managed bridge and Relay runtime.
To detect installed platforms and choose one interactively (including Codex, Claude Code, and Pi), run this in your computer's terminal:
npx @p697/clawket@latest pair choose
The pairing home shows this command and its Copy action directly below the platform list, under Or detect automatically. Scan and expandable code entry are available on the same page, and the scanner can also read a QR code from your photos; manual codes require selecting their platform first. Platform-specific onboarding defaults to terminal scanning for Codex, Claude Code, Pi, and local models; OpenClaw and Hermes default to an agent message. You can switch methods and enter a pairing code instead of scanning. pair choose is interactive, accepts no flags, and is not a Preview command.
For direct pairing on your local network:
clawket pair local
To select a backend explicitly, use --backend with openclaw, hermes, codex, claude-code, pi, or local-model. Use clawket status, clawket doctor, and clawket logs to inspect your connection.
status summarizes all saved Agent connections; add --verbose for paths and capabilities. Use clawket logs --backend codex --last 10m --follow (or another backend) for live troubleshooting. Default lifecycle/reset commands retain the OpenClaw/Hermes service scope; use clawket codex reset, clawket claude-code reset, or clawket pi reset with the original pairing options for those Agents. Reset retains session history.
Use Node.js 22.x and npm for this checkout. iOS development requires macOS and Xcode; Android development requires Android Studio and its SDK.
npm ci
# Generate only the platform you are developing (Android also works without Xcode):
npm exec --workspace apps/mobile -- expo prebuild --platform ios
npm run mobile:dev:ios
# Or, generate Android and run it:
npm exec --workspace apps/mobile -- expo prebuild --platform android
npm run mobile:dev:android
Optional public configuration is documented in apps/mobile/.env.example. Copy it to apps/mobile/.env.local to configure your build. Provider secrets belong on the server, never in EXPO_PUBLIC_* values.
Community Debug and Release builds work without analytics, billing or speech configuration. Leave CLAWKET_OFFICIAL_BUILD unset; the production/testflight EAS profiles are for official Clawket distribution. Use your own Apple signing team for a physical device and your own EAS project for cloud builds; see self-hosting. iOS also requires CocoaPods, and Android requires JDK 17 with ANDROID_HOME set.
npm run mobile:config:show
npm run mobile:config:check
Local-model connections are also available for llama.cpp, Ollama, and OpenAI-compatible servers. See the local-model guide for setup and current limitations.
In Relay mode, the Registry handles pairing and the Relay forwards live WebSocket traffic between the app and your Bridge. Your agent continues to run on your own machine. In direct mode, the app reaches your backend endpoint over LAN, Tailscale, or a custom URL without Relay infrastructure.
You can build the mobile app and operate your own services. OpenClaw and Hermes share the Relay implementation but use separate services, storage, and pairing credentials.
Relay forwards traffic without persisting message content. The app keeps a local message cache; deleting a connection clears its cache. Your agent backend and model provider retain their own data-handling policies.
Source builds can leave these integrations unconfigured or configure their own services:
| Component | Purpose |
|---|---|
| PostHog | Basic usage and diagnostics; analytics events exclude conversation text and prompts. Leave analytics configuration empty to disable it. |
| RevenueCat | Store subscription management. Unconfigured source builds skip billing and unlock Pro features. |
| Alibaba Cloud speech service | Transcribes audio when you use cloud voice input. Configure the independent speech endpoint to enable it; provider keys stay on the server. |
These components are separate from Agent connectivity. See the configuration example and speech service guide for details. Store builds use their configured integrations; see the Privacy Policy.
| Path | Purpose |
|---|---|
apps/mobile | Expo / React Native app |
apps/bridge-cli | Published @p697/clawket CLI |
apps/relay-registry | Pairing Registry Worker |
apps/relay-worker | WebSocket Relay Worker |
apps/speech-worker | Independent cloud transcription service |
packages/agent-protocol | Shared backend contracts and capabilities |
packages/bridge-core | Pairing, configuration, and service helpers |
packages/bridge-runtime | Bridge runtimes |
packages/relay-shared | Shared Relay protocol and types |
Read CONTRIBUTING.md. Run the repository's required checks before proposing code changes:
npm run check:required
Live-service and release integration checks have separate prerequisites; see the contributor and workspace documentation. Report security issues as described in SECURITY.md.
Unless a subdirectory states otherwise, this repository is licensed under AGPL-3.0-only.
Bridge upgrades preserve existing pairing. See Bridge updates for unified update support, runtime verification and manual deployment exceptions.