manim-cli
v0.1.2
Published
Production-ready npm wrapper for Manim Community with agent workflows and multi-provider TTS
Maintainers
Readme
manim-cli
manim-cli is an npm-distributed wrapper for Manim Community Edition with:
- Manim-compatible render commands
- Agent-oriented pipeline commands for Codex/Claude Code
- Local-first multi-provider TTS with Kokoro as the default provider
- OpenAI cloud fallback and managed voice/profile configuration
Status
This repository contains:
- the npm CLI
- agent workflow and proposal gating
- managed runtime bundle installation
- local-first TTS orchestration
- runtime packaging scripts and fixture tests
Managed Runtime
manim-cli is designed for a zero-manual-install setup flow on supported platforms:
npm install -g manim-climanim-cli setupmanim-clidownloads and verifies a platform bootstrap bundle, then creates the managed runtime locally at its final install path
Supported runtime targets:
darwin-arm64darwin-x64linux-x64linux-arm64win32-x64
Important detail:
- Manim Community no longer requires an external
ffmpegbinary for rendering itself. manim-clistill bundlesffmpegandffprobebecause the CLI uses them for captions, muxing, and final composition.
Runtime bootstrap bundles are published as GitHub Release assets and referenced by the release-generated runtime/runtime-manifest.json. That file is intentionally empty in the repo until real platform assets are built and published. See runtime/README.md for the bootstrap-bundle format, provisioning commands, release asset naming, rollback policy, and clean-machine acceptance checklist.
GitHub automation
This repo ships with:
- CI for
npm ci,npm run build, andnpm test - Runtime Release for publishing managed runtime bundles
- OpenAI PR Review for automatic pull request analysis using the OpenAI API
- OpenAI Failure Review for automatic failed-build analysis using the OpenAI API
Setup required before these workflows will work:
- Add
OPENAI_API_KEYto GitHub Actions secrets - Optionally add a repository variable
OPENAI_REVIEW_MODEL - Re-run the workflow
These workflows use the OpenAI Responses API and default to gpt-5-codex, which OpenAI documents as a coding-optimized model available in the Responses API. OpenAI also documents Codex SDK/API use in CI/CD contexts and shows Responses API examples that read OPENAI_API_KEY from the environment: Code generation guide, GPT-5-Codex model, API libraries/auth setup.
