Releasesv0.18.0
v0.18.0 — Batteries included
This is the release note as published, not a rewrite of it. Release notes are published in the language they were written in.
"Batteries included"
Added
pip install 'chimera-agent[full]'— one command installs every non-GPU feature: messaging adapters, MCP, document ingestion (PDF/Word/Excel→text), media download (yt-dlp), local speech-to-text (faster-whisper), data analysis (pandas/scikit-learn) and charts (matplotlib/seaborn/plotly), and YouTube transcripts. The GPU-heavy extras (imagegen-local,train) stay opt-in.- Batteries-included Docker image — the official image now builds from
[full]and installsffmpeg, so vision, audio (speech-to-text + text-to-speech), video, documents, data & charts, and the browser all work out of the box in the container.
Documentation
- A complete-beginner "🧰 What Chimera can do — and how to switch each thing on" section in all 7 READMEs (English, Português, Español, Français, Deutsch, 中文, 日本語): a point-by-point table of every ability with exactly which extra/key it needs and the command to try it, a one-line "turn everything on", and a six-step first-time setup.
pip install -U 'chimera-agent[full]' · Apache-2.0