A proxy server for OpenAI-compatible chat completion endpoints, built using Apache Pekko. Handles proxying to remote providers and managing local llama-cpp instances.
radisson --config=config.yaml
See config.example.yaml for available configuration options.
- Scala 3.7
- sbt 1.12
- (Optional) llama-cpp for local backends
# Compile the project
sbt compile
# Run tests
sbt test
# Build a fat JAR
sbt assembly
# Build a GraalVM native image executable
sbt nativeImage
# Running in dev mode
sbt run -- --config=config.yaml
# fix lint
make fix-lint