Cut your OpenClaw costs by up to 70%
Manifest intercepts your query and passes it to the most cost-effective model.
All your requests are going to the same model
OpenClaw just routes everything to the same model, regardless of the complexity of the query. You are summoning big models for tiny tasks.
Reduce your OpenClaw costs
Manifest analyzes each query locally in less than 2ms and picks the cheapest model that can handle it. 300+ models across every major provider, one routing layer.
/ Budgets Limits
Stop runaway costs before they happen.
Set a spending limit on a defined period. Manifest emails you when the threshold is crossed and can block routing until the period resets.
/ Add custom providers
Add any OpenAI-compatible provider.
Connect any provider that exposes an OpenAI-compatible endpoint, add the models you want, and they join your routing tiers like any other.
/ Model Fallbacks
Your agent never sees a failure.
Up to 5 fallback models per routing tier. If the primary is rate-limited or down, Manifest retries the next one. Fall back across providers or to a local model when budget runs out.
/ Spend Tracking
Know exactly where your money goes.
Every message shows its cost, input tokens, output tokens, and cache reads. Filter by agent, model, or time range.
| Model | Pricing / 1M tokens | Tokens / last 7 days |
|---|---|---|
| Loading... | ||
| Model | Pricing / 1M tokens | Tokens / last 7 days |
|---|---|---|
| Loading... | ||
Stop wasting money on the wrong model
One plugin. All your models. Easy to install, easy to trust.
Get startedFrequently asked questions
Is your API compatible with OpenAI?
Yes. Our endpoints and schemas are drop-in compatible with the OpenAI API, so you can integrate in minutes.
Do you retain my data?
We don't store your prompts or responses. We only log metadata (model used, token count, cost). In local mode, nothing leaves your machine.
How does routing choose a model?
Each request goes through a scorer that checks the prompt length, technical keywords, tool usage, and other signals to classify it into one of four complexity tiers (simple, standard, complex, reasoning). Then it picks the cheapest model assigned to that tier from your connected providers.
Can I use Manifest with free models?
Yes. We support free models from OpenRouter and other providers. You can mix free and paid models in your routing. Simple tasks go to free models, complex ones go to paid ones.
What happens if my primary model fails?
You can add up to 5 fallback models per tier. If a provider returns an error or times out, the request goes to the next model in the chain. Your agent keeps working.
What's the difference between cloud and local mode?
Same features, same routing. Cloud routes through app.manifest.build. Local runs on your machine, nothing leaves your network, and you can route to local models via Ollama.
Messages routed through Manifest
250,000+