A command-line AI assistant that provides contextual awareness by accessing your local filesystem and web content. Built using Kotlin, LangChain4j, and MCP.
- Terminal-based AI chat CLI with markdown support
- MCP
- Persistent chat memory
Create a config file at ~/.contexta/config.properties with the following:
baseUrl=<openai compatible url>
apiKey=<your api key>
modelName=<model name>The project uses Amper build system. To build:
./amper buildRun the application:
./amper runOr directly with the generated JAR:
./amper package && java -jar build/tasks/_cli_executableJarJvm/cli-jvm-executable.jar- Type your query and press Enter to send it to the AI
- Type
/exitto quit the application
- cli: Command-line interface implementation with terminal UI
- integration: Core logic for model interaction, tool integration, and configuration
MIT