manifest
Model Router for OpenClaw. Save up to 70% by routing requests to the right model. Choose LLM fallback to avoid API rate limits, set thresholds and reduce token consumption.
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/brunobuddy/manifest-buildManifest — LLM Router & Observability for OpenClaw
Manifest is an OpenClaw plugin that:
- Routes every request to the most cost-effective model via a 23-dimension scoring algorithm (<2ms latency)
- Tracks costs and tokens in a real-time dashboard
- Sets limits with email alerts and hard spending caps
Source: github.com/mnfst/manifest — MIT licensed. Homepage: manifest.build
Security & Privacy
TL;DR — The plugin registers Manifest as a standard OpenAI-compatible provider and exposes three read-only agent tools. It does not export telemetry or make background network calls. When you select
manifest/autoas your model, OpenClaw routes requests through the Manifest backend — the same way it routes to any other provider like Anthropic or OpenAI. In local mode, all data stays on your machine and no API key is needed.
What the plugin does
- Registers a provider — adds
manifestas an OpenAI-compatible provider with theautomodel - Injects config — writes provider entry to
~/.openclaw/openclaw.jsonand auth profiles (standard plugin registration, reversed on uninstall) - Exposes 3 read-only tools —
manifest_usage,manifest_costs,manifest_health(query your own usage data via the Manifest API) - Registers
/manifestcommand — shows connection status
What the plugin does NOT do
- Does not export telemetry, traces, or metrics — the plugin has no OTLP or telemetry code
- Does not make background or periodic network calls — network calls happen only at startup (connection verification) and when agent tools are invoked by the user
- Does not change your default model —
manifest/autois added to the allowlist only, you must switch to it manually
How routing works
When you manually select manifest/auto as your model, OpenClaw sends requests to the Manifest backend's /v1/chat/completions endpoint — the same way it sends requests to any provider (Anthropic, OpenAI, etc.). The backend picks the optimal model based on conversation complexity. This is standard OpenClaw provider behavior, not a special plugin data flow.
Credential storage
- Cloud mode (
manifest-model-routerplugin): API key provided viaopenclaw providers setup manifest-model-routerorMANIFEST_API_KEYenv var. The key authenticates with the Manifest backend — standard provider auth. - Local mode (
manifestplugin): auto-generated key stored in~/.openclaw/manifest/config.jsonwith file mode0600. No external service contacted.
Local mode
All data stays on your machine. The embedded server runs locally and no external calls are made.
Install Provenance
openclaw plugins install manifest installs the manifest npm package.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-brunobuddy-manifest-build": {
"enabled": true,
"auto_update": true
}
}
}Related Skills
webmcp
This skill should be used when browsing or automating web pages that expose tools via the WebMCP API (window.navigator.modelContext). It teaches agents how to discover, inspect, and invoke WebMCP tools on websites instead of relying on DOM scraping or UI actuation.
Startup Info
Skill by brunobuddy
hide-openclaw-banner
Suppress the OpenClaw CLI banner message. Use when the user says "hide openclaw banner", "disable openclaw banner", "suppress openclaw banner", "remove openclaw banner", "openclaw banner is annoying", "turn off the lobster", or wants to stop seeing the OpenClaw startup tagline on every command invocation.