Welcome to taoCode
taoCode is an AI coding agent that works in your terminal. It can read your project, edit files, and run commands. Its built-in TAO provider connects those local tools to models served through Bittensor gateways.
Start here
Section titled “Start here”| If you want to… | Read… |
|---|---|
| Set up taoCode on your machine | Installation |
| Complete your first task | Quickstart |
| Connect an inference backend | Gateways |
| Change the model for a task | Selecting a model |
| Look up a command or file location | CLI reference · Configuration |
Understand the workflow
Section titled “Understand the workflow”- Start locally. Open a terminal in the project you want to work on.
- Choose inference. Connect a gateway and select one of its model references.
- Give the agent a task. Describe the change and how you expect it to be checked.
- Review the result. Inspect the diff and test output before keeping the changes.
The agent and tools run on your machine. Model requests send prompts and relevant context to the selected gateway. Read Privacy & decentralization to understand that boundary.
A few terms
Section titled “A few terms”Provider is the model integration inside taoCode. Gateway is the API backend it connects to. Subnet is a specialized network within Bittensor. A model reference, such as tao/chutes/Qwen/Qwen3.8-27B-TEE, identifies a provider, gateway, and model.
Continue with Installation or explore the architecture.