1
Here's what's available:
- Google Gemini - 5 models including gemini-2.5-pro and gemini-2.5-flash. Up to 250K tokens per minute across all models. The pro model has a 1M context window on the free tier.
- Cohere - command-a-03-2025 and command-a-reasoning-08-2025. 1,000 calls per month, 256K context.
- Kilo Code - 4 models including Qwen 3.6 Plus, Nemotron 3 Super 120B, and Step 3.5 Flash. Around 200 requests per hour. Some support image and video input.
The whole point is to get started without spending anything. Connect one or two free providers, set up a routing config with fallbacks, and you already have a working setup. If Gemini hits its rate limit, Manifest falls back to Cohere or Kilo Code automatically.
More ready-to-setup free models are coming. hey are all listed here: https://manifest.build/free-models
We're still in beta and actively trying to understand how people use this. What does your setup look like? What providers are you using? If you run into anything weird or have feedback, we want to hear it.
What’s one specific thing you wish your OpenClaw agent could do today, but can’t?
Not vague stuff like “pay for things.” I mean which concrete use case ?
For example:
- “Automatically renew my AWS credits if usage drops below $100 and pay with a virtual card.”
- “Find the cheapest nonstop flight to NYC next month, hold it, and ask me before paying.”
How do teams then handle monitoring, bugs, scaling, and reliability with many users?
I’m curious whether any Ops teams here (RevOps, Finance Ops, BizOps, etc.) are using ChatGPT regularly for internal, enterprise work, especially in environments where information is spread across many tools.
If so, what does your current workflow look like? Are people copy-pasting data or documents, asking IT to build custom GPT apps, relying on custom agents, or using other internal solutions? What feels painful or inefficient today, and are you actively looking for better ways to handle this?