deltafin
Run Kimi K3 locally with an OpenAI-compatible API server for chat and coding agents.
free(Free tier)799 github stars
About deltafin
Deltafin enables running the full Kimi K3 model on a single device, targeting local deployment without cloud dependencies.
It provides an OpenAI-compatible API server, allowing integration with existing tooling, chat interfaces, or coding agent frameworks that expect the OpenAI REST interface.
The project is built around local AI execution and focuses on compatibility and self-hosted usability.
Key Features
- Runs the full Kimi K3 model locally on a single device
- Exposes an OpenAI-compatible HTTP API for chat and coding agent use
- Supports local LLM inference without requiring cloud services
- Designed for integration with agent frameworks expecting OpenAI-style endpoints
Pros & Cons
Pros
- • Enables fully offline, private execution of Kimi K3
- • Simplifies adoption by reusing standard OpenAI client libraries and tooling
- • Lightweight local deployment target for students and small teams
Cons
- • No license information is provided, creating uncertainty about permitted usage and redistribution
- • Lack of documentation on hardware requirements or performance benchmarks makes deployment planning difficult
- • No website or official support channels limit discoverability and troubleshooting