Chalkboard
github.com
What it is
Chalkboard is free, MIT-licensed software that turns a prompt into a narrated explainer video and exports an MP4. It supports 16:9 explainers, vertical short-form reels, and cinematic documentaries. Cinematic mode researches a topic and places cited sources in the script. Productions can be dubbed into multiple languages in one run, with subtitles set independently from narration language. Script generation works with Anthropic, OpenAI, or local Ollama; narration can use Piper, OpenAI, or ElevenLabs. Chalkboard creates hand-drawn diagrams, supports SVG artwork, and can generate images through OpenAI’s gpt-image-2. Subtitles and mood-matched music are enabled by default, with Jamendo or user-supplied tracks available. The project provides a command-line interface, library, HTTP API, and web studio, and runs self-hosted. Ollama and Piper support local generation at zero marginal cost per video. Generated images require an OpenAI API key and may incur costs.
Spec sheet
- Purpose
- Chalkboard turns a prompt into a narrated, whiteboard-style explainer video exported as an MP4.Read 9 Oct 2026 · github.com
- License
- Chalkboard is released under the MIT License.Read 9 Oct 2026 · github.com
- AI providers
- Script generation supports Anthropic, OpenAI, and local Ollama; narration supports Piper, OpenAI, and ElevenLabs.Read 9 Oct 2026 · github.com
- Audio
- Subtitles and mood-matched background music are enabled by default, and users can choose Jamendo music or provide their own track.Read 9 Oct 2026 · github.com
- Cost limits
- Generated images require an OpenAI API key and incur costs; the README says image generation can be disabled or capped per video.Read 9 Oct 2026 · github.com
- Languages
- A production can be dubbed into multiple languages in one run, with subtitles in a language different from the narration.Read 9 Oct 2026 · github.com
- Research
- The Cinematic mode researches topics and injects cited sources into the script to ground the narration.Read 9 Oct 2026 · github.com
- Security detail
- The README says API keys are read from the server environment and the web UI never sees them.Read 9 Oct 2026 · github.com
- Self-hosting
- The project runs on the user's machine and can use Ollama and Piper for local script generation and narration at zero marginal cost per video.Read 9 Oct 2026 · github.com
- Video formats
- It supports 16:9 explainers, vertical short-form reels, and research-backed cinematic documentaries.Read 9 Oct 2026 · github.com
- Visuals
- It creates hand-drawn diagrams and can generate images with OpenAI's gpt-image-2; SVG art is also supported.Read 9 Oct 2026 · github.com
- Ways to use
- The project provides a command-line interface, a library, an HTTP API, and a web studio.Read 9 Oct 2026 · github.com
Plans and prices
Strong points and limits
Strong points
- Free MIT-licensed software
- Multiple video formats and languages
- CLI, library, API, and web studio
- Local providers can avoid per-video costs
- Cited research mode for documentaries
Limits
- Generated images can incur OpenAI costs
- Image generation requires an API key
- Self-hosted setup is required
Who it suits
It suits creators who need narrated whiteboard, reel, or documentary videos from prompts. Self-hosting and local Ollama or Piper options fit users who want local generation.
Verdict
Chalkboard combines prompt-based scripting, drawing animation, narration, research, and export in one self-hosted project. Its local provider support is useful, but image generation has separate OpenAI requirements and costs.
Where Specifiction places it
- Reverse Engineering Tools#21 of 737.7
- Whiteboard Animation Software#14 of 267.8