The Moltbot Phenomenon
Moltbot (previously Clawdbot) has exploded onto the scene as one of the fastest-growing open-source AI agents on the planet.
In just 48 hours, it accumulated over 50,000 GitHub stars, and drove massive Mac Mini sales.
We greatly respect creator Peter Steinberger for navigating this incredible growth while dealing with a necessary name change from Clawdbot to Moltbot due to a trademark request from Anthropic, dealing with bad actors hijacking his handle, and managing yesterday's Product Hunt co-launch with Kilo, all while the project explodes in popularity.
But here's the challenge that Moltbot users face: running an AI agent gets expensive, fast. Many users try running models locally to save on API costs, but this creates a new problem:
Serious GPU Hardware Required
Running the latest open-weight models like GLM 4.7 locally at 4-bit quants requires serious GPU hardware with at least 40GB ($4000+) and 200GB of RAM ($1000+).
For most users, local inference isn't the solution, you need cost-effective API access to the models.
Proactive Agents = $$$
Unlike reactive AI tools where you pay per query, proactive agents like Moltbot are constantly working in the background:
Monitoring websites for changes
Checking news sources for relevant updates
Scanning your emails and calendar
Running scheduled tasks and reminders
Executing multi-step workflows
All of this consumption adds up quickly when you're paying standard API rates. Users report spending significantly more than expected within their first weeks of use.
How Kilo Gateway Solves This
We're making it affordable to run Moltbot by giving you access to exclusive free models and cost-effective pricing on frontier models.
Access Free and Cost-Effective Models
Kilo Gateway offers free and cost-effective models. Models like GLM-4.7 and MiniMax M2.1 are exclusively free on Kilo right now, perfect for many agent tasks.
When you need more powerful models for complex reasoning, get easy access to frontier models like Claude Opus 4.5, GPT-5.2, and Gemini with pay-as-you-go pricing, without markup.
Or use a Kilo Pass to get extra free usage on top of your AI subscription.
Getting Started: Two Ways to Use Kilo Gateway in Moltbot
Option 1: Manual Configuration (Available Now)
You can start using Kilo Gateway with Moltbot right now by adding it as a custom provider:
Get your Kilo API key from app.kilo.ai/profile
Open your Moltbot configuration:
- Navigate to your ~/.clawdbot folder
- Edit the clawdbot.json file
- Add the Kilo Gateway configuration:
"kilocode": {
"baseUrl": "https://api.kilo.ai/api/openrouter/",
"apiKey": "YOUR-API-KEY-FROM-app.kilo.ai/profile",
"api": "openai-completions",
"models": [
{
"id": "anthropic/claude-opus-4.5",
"name": "Anthropic: Claude Opus 4.5"
},
{
"id": "minimax/minimax-m2.1:free",
"name": "Minimax: Minimax M2.1"
},
{
"id": "zhipuai/glm-4.7:free",
"name": "GLM-4.7 (Free - Exclusive to Kilo)"
},
// {
// "id": "any-other-models-you-want",
// "name": "See https://api.kilo.ai/api/openrouter/models for full list"
// }
]
}
- Save and restart Moltbot - You'll now see Kilo Gateway models available in your provider selection
Option 2: Setup Wizard Integration (Coming Soon)
We've opened a pull request to merge Kilo Gateway directly into Moltbot's setup wizard. Once merged, you'll be able to select Kilo Gateway during initial configuration with no manual JSON editing required.
Why This Matters for Agents
With Kilo Gateway, you get access to very cost-effective models---including exclusive free ones like GLM-4.7 and MiniMax M2.1 for routine tasks, and frontier models like Claude Opus 4.5 and GPT-5.2 for complex reasoning. Use free models for the bulk of agent work, and only pay for premium models when you actually need them.
Ready to Run Moltbot Without Breaking the Bank?
Sign up at kilo.ai
Follow the setup steps above to add Kilo Gateway to your Moltbot configuration
Start running your agent with exclusive access to free models, cost-effective and frontier models
Top comments (0)