Skip to content

Releases

AgentFlight 0.4.0

AgentFlight 0.4.0 Latest
Latest

Choose a tag to compare

View all tags
@AkiCagkan AkiCagkan released this 13 Oct 06:48
v0.4.0
f2ee66e

First public release of AgentFlight: write LLM agents in plain Python, watch every call live, step in, replay any step, and test them with evals, from a panel inside VS Code / Cursor (or the browser).

Install

pip install agentflight-0.4.0-py3-none-any.whl          # or: pip install -e . from a clone (Python 3.11+)
code --install-extension agentflight-vscode-0.4.0.vsix  # VS Code; Cursor: cursor --install-extension ...

Then run agentflight watch and start any program that uses AgentFlight; it shows up in the panel.

Assets

File
agentflight-0.4.0-py3-none-any.whl the Python library and the agentflight command
agentflight-vscode-0.4.0.vsix the VS Code / Cursor extension
agentflight-jetbrains-0.4.0.zip the JetBrains plugin (experimental: built against the IDE APIs, not yet tried in a running IDE)

Highlights

  • Agent / @llm: typed answers (Schema, Pydantic, plain types) with automatic format fixes, tools and MCP servers, parallel calls, human approvals, budgets, memory, async methods.
  • Providers: Claude Code (subscription) and the Anthropic API, Sign in with ChatGPT and the OpenAI API, Gemini, Ollama, LM Studio, and an offline fake model.
  • Every run is recorded and can be replayed from any step; evals with thresholds and a model judge.
  • 118 offline tests plus browser and VS Code extension tests; CI on Linux, Windows and macOS with Python 3.11-3.13.

See the README and the changelog.

Assets 5
Loading