AI-powered CLI for API testing: generate test data and scenarios from OpenAPI, analyse failures, and inspect traffic.
Download the binary for your platform from the releases page.
chmod +x jarvis
sudo mv jarvis /usr/local/bin/On Windows, put jarvis.exe on your PATH. To build from source, see the setup guide.
jarvis setup
jarvis gen generate-testdata --spec api.yaml --count 10
jarvis gen generate-scenarios --path specs/openapi/api.yaml
jarvis analyze analyze-failures --limit 5
jarvis proxy --record
jarvis certificate --cert-dir ./certsGeneration and failure analysis call one LLM through pkg/engine/llm. Set LLM_PROVIDER to openai, anthropic, google, or ollama. If it is unset, Jarvis checks OPENAI_API_KEY, then ANTHROPIC_API_KEY, then GOOGLE_API_KEY, and otherwise uses Ollama.
| Provider | How to authenticate | Default model |
|---|---|---|
| OpenAI | OPENAI_API_KEY |
gpt-4o-mini |
| Anthropic | ANTHROPIC_API_KEY |
claude-3-5-sonnet-20241022 |
| Google Gemini | GOOGLE_API_KEY |
gemini-1.5-flash |
| Ollama | local server (OLLAMA_HOST, default http://localhost:11434) |
llama3.2 |
LLM_MODEL overrides the default model. LLM_API_KEY is accepted in place of the provider-specific key, and LLM_BASE_URL sets a custom endpoint for OpenAI-compatible APIs or for Ollama.
export OPENAI_API_KEY="sk-..."
export LLM_PROVIDER=openai
export LLM_MODEL=gpt-4o-miniProvider setup, model choice, and troubleshooting: LLM integration and the provider quick start.
- Commands
- AI features: test data, scenarios, and failure analysis
- Traffic inspector: proxy, TLS, and OpenAPI validation
- Workflows: examples carried over from the previous README
- Setup and configuration
- Further examples
git clone https://github.com/dipjyotimetia/JARVIS.git
cd JARVIS
go mod download
go build -o jarvis
go test ./...MIT. See LICENSE.
