animus

tryanimus.app

Turn any question into a narrated explainer video.

animus is an interactive coding agent for Manim videos. Think Claude Code, but in the cloud and pointed at explainer animations.

• Describe a topic. The agent researches it and settles a scene plan with you.
• It writes and renders Manim in a persistent cloud sandbox.
• It reads its own errors and repairs until the scene is clean.
• Narration is synthesized during the render, so the animation times itself to the speech instead of the other way around.

That render, diagnose, repair loop driven conversationally is the product. Not one shot codegen.

What it does

• Researches before it claims. Any specific number, date, quote or recent event gets a web search first. A wrong fact in a finished video is treated as a defect, not a rounding error.
• Plans with you, not at you. The agent asks when a decision could reasonably go two ways, then proposes an ordered scene list for you to approve or send back. No production work starts before you agree.
• Narrates itself. Scenes are written as manim-voiceover scenes against ElevenLabs. Each beat is wrapped in its narration, so run_time follows tracker.duration and the timing cannot drift. A music bed is mixed underneath after the render.
• Takes any subject. Manim is an animation engine, not a maths tool. Biology, history, distributed systems and monetary policy are as much in scope as calculus.
• Repairs its own work. Render fails, agent reads the traceback, edits the file, renders again. You watch it happen and can interrupt at any point.
• Ships somewhere. Finished videos download as mp4 or publish to a permanent unlisted link with a branded page and real link previews.

How it works

Each user turn is one AI SDK tool calling loop, streamed to the browser over SSE and abortable mid flight.

What holds the design together.

• The turn is ephemeral. If it crashes or you stop it, you just send another message. Nothing durable is lost.
• The sandbox is the state. One Daytona sandbox per conversation, suspended between turns and resumed on the next message. The Manim project lives there across turns.
• So a video is never a one shot artifact. Come back three turns later and ask for a slower second act, a different colour, one more scene: the agent edits the same files, re renders, and hands you a new cut.

What the agent can reach for.

Tools	What they do
writeFile editFile readFile listFiles	the Manim project in the sandbox
runCommand	a shell in the sandbox
renderScene	render, mix the music bed, deliver the mp4
webSearch webFetch	research through Exa
askUserQuestion finalizeVideoPlan	pause the loop until you answer

Sandboxes boot from a prebaked snapshot with Manim, ffmpeg and LaTeX already installed, so there is no per turn cold install.

Stack

Layer	Choice
Runtime, package manager, tests	Bun, Vitest
Monorepo	Turborepo, source first (no per package build)
Web	React 19, Vite, React Router, Tailwind v4
API	Hono, long running container
Agent	Vercel AI SDK, Claude via Amazon Bedrock, BYOK for Anthropic, OpenAI and Google
Sandbox	Daytona
Data	Postgres (Neon) with Drizzle, Cloudflare R2 for video
Auth	Better Auth (magic link, GitHub, Google)
Narration	manim-voiceover, ElevenLabs
Quality	Ultracite and Biome, Knip, strict TypeScript

• Metered per component in integer micro USD.
• Bring your own model or narration key and that half runs unmetered on your own account.

Run it locally

cp .env.example .env   # fill in what you have; the file has notes
docker compose up -d   # Postgres on :5432
bun install
cd packages/db && bun run db:migrate && cd ../..
bun run dev            # web on :5173, api on :8787

Env notes.

• Required even if you never render: ELEVENLABS_API_KEY and the four R2_* vars, at any non empty value. The API validates its env at boot. Known rough edge.
• Genuinely optional: DAYTONA_API_KEY and EXA_API_KEY, resolved only when a turn needs the sandbox or web research.
• Without RESEND_API_KEY, magic links print to the API console instead of being emailed, which is enough to sign in.
• Narration needs a paid ElevenLabs tier. The free tier caps at 10k characters a month and blocks synthesis from datacenter IPs, which is where the sandbox runs.

Commands

bun run dev        # turbo dev (all apps)
bun run build      # turbo build (web only)
bun run typecheck  # tsc across every workspace
bun run check      # ultracite check, repo wide lint and format
bun run knip       # unused files, deps and exports
bun run test       # the test suite (not `bun test`)

Layout

apps/
  web/    React SPA: landing, auth, studio, settings, share pages
  api/    Hono server: auth, conversations, the streaming agent loop, share endpoints
packages/
  core/   shared Zod contracts, pricing, share card and link preview builders
  agent/  the loop, prompts, Manim and sandbox tools, the bundled manim-video skill
  auth/   Better Auth
  db/     Drizzle schema, client and migrations

Contributing

Contributions are welcome.

• CONTRIBUTING.md — setup, the four checks every change has to pass, and the conventions the codebase holds to.
• CLAUDE.md — the full decision record, and why things are the way they are.

License

MIT. See LICENSE, and NOTICE.md for bundled third party assets.