Deploy a Python function to Animica Python Cloud and you get a public HTTPS endpoint plus 80% of every ANM someone spends running it — metered per call and settled the moment it finishes. Your deployment is anchored on-chain; execution runs off-chain in a sandbox with no network. The AI is free, Animica Workers are free, and there is no card to start.
You choose the prompt, the cadence and the destination. Workers no longer need a subscription — every account gets them free, with a 15-minute cadence floor and fair-use caps. Repeated consecutive failures disable a Worker automatically (five by default, per Worker), and Emergency Stop is one click.
One ladder for everything — Cloud, Workers and the CLI: Free $0 · Starter $4.99 · Pro $14.99 · Operator $39.99 · Business $99.99 — the whole ladder ↓
A subscription doesn’t buy the intelligence. It buys the alarm clock, the delivery and the log. Workers run on the same free Animica AI you are about to use. Nothing above is a trial and nothing below is a preview — the free AI is not a quota you burn down. Here it is, running right now:
Free, no key, no signup — the same AI your Workers and agents run.
Write a function. Deploy it. You get a public HTTPS endpoint and 80% of every ANM a caller spends running it. The other 20% runs the platform. No servers, no containers to manage, no invoices to chase — settlement happens on every single execution.
A plain Python function. The whole interface is one entrypoint:
def main(request):
name = request.get("name", "world")
return {"message": f"Hello {name}"}Need more? animica.ai.infer() for AI, animica.call() to buy another app’s output, animica.state to remember things — each one a capability you declare.
One command, or the in-browser editor:
pip install animica
animica cloud deploy handler.pyYour source hash and artifact hash are anchored on-chain in a real Animica deploy transaction. Execution itself runs off-chain, in a locked-down sandbox with no network.
Every call is metered — CPU, memory, AI tokens — priced in ANM and settled the moment it finishes. Your share lands in your balance immediately and withdraws on-chain to your wallet.
Publish to the marketplace and your app gets its own page people can find and share. You keep 80%.
Deploy Python Explore Apps Read the docs →
Honest about the architecture: Animica anchors your deployment on-chain and executes it off-chain. Consensus does not run arbitrary Python — the chain records which code was deployed, and the network pays for running it.
An agentic coding assistant that reads your files, edits them, runs commands and checks its own work — the way Claude Code and Codex do. No API key, no wallet and no GPU. Inference comes from the same miner network that serves this page.
Full CLI guide → — install, the four approval modes, /swarm, AGENTS.md, and what a paid plan lifts.
# a venv keeps it out of your system Python
python3 -m venv .venv && source .venv/bin/activate
pip install animica
animica chat
On Windows the activate line is .venv\Scripts\activate. Prefer one command that makes the venv for you? curl -fsSL https://animica.dev/install.sh | sh — and you can read it first; it is about 120 lines and needs no sudo.
animica chat
Slash commands: /help, /mode, /swarm, /sessions, /resume, /exit. Sessions are saved on your machine, so animica chat -c picks up the last one and --session <id> picks any of them. Drop an AGENTS.md in the repo and it follows it.
# reads, greps, edits, runs tests — asks before each change animica chat "find the N+1 query in this repo and fix it" # pipe anything in: a stack trace, a failing build, a diff pytest 2>&1 | animica chat "why is this failing?"
Tools are on for every message — there is no separate agent mode. Ship it with animica deploy handler.py. Or run several agents at once: type /swarm audit this codebase for auth bugs inside a session. It plans the work into parallel subtasks, runs them together, has a separate agent try to refute each result, and merges only what survived.
Write a Python function, deploy it to Animica Cloud, and get paid when people call it — without leaving your shell.
# one file, one command animica cloud login animica cloud init # writes a handler.py to start from animica deploy handler.py # validates, uploads, anchors on-chain, activates
Then animica cloud invoke, logs, status and earnings do what they say. animica deploy is a shortcut for animica cloud deploy.
Every session runs in one of four approval modes. Switch any time with /mode, or start in one with --permission-mode. The mode is always on screen, because an agent that auto-approves shell commands without saying so is indistinguishable from a careful one until something is already gone.
Reads only. Writes, shell and network are refused — it will not even ask.
The default. Asks before writing a file, deleting, running a shell command or fetching a URL.
Edits apply as it goes; shell and delete still ask. What most coding sessions want.
Everything, unattended. Switching into it mid-session asks you to type auto to confirm.
No card, no key, no wallet.
The CLI is part of every paid plan — not a separate subscription. Paste the licence key from your plan into animica chat → /licence <key>.
Running a GPU? animica up serves this network and earns ANM — the CLI you just installed is also the miner.
Everything on this page stays free — chat, the media and GPU studios, the swarm builder, the /v1 API, and Animica Workers, which no longer need a subscription at all. A plan does not unlock the AI; it raises how much you can run.
The ladder is executions, compute and concurrency on Python Cloud — how much your deployed functions may consume in a month. Every tier earns the same 80% developer share on what your users pay.
Deploy your first function
500 executions / mo
◦ Nothing above ever moves behind a plan.
Deploy now →Publish and earn
100,000 executions / mo
Priority and depth
1,000,000 executions / mo
Production workloads
10,000,000 executions / mo
Dedicated capacity
Custom quotas & reserved capacity
Checkout connects your Animica wallet — the same sign-in the developer tools use. Billed monthly in USD through PayPal. Change or cancel any time: on a downgrade, resources over the new limit keep their configuration and simply stop running. Nothing is deleted because billing stopped. Plans buy platform capacity; they are not ANM and not an investment. What your users pay to run your functions is separate, is denominated in ANM, and is 80% yours on every tier.
No key, no signup. Every job is dispatched to the Animica GPU-miner network and comes back with a post-quantum-signed receipt — no model ever runs on our servers. Advanced multi-scene video and image-to-video are built in. Your uploaded images stay private, seen only by the miner that renders your clip.
Describe an image and a GPU miner will render it. PNG out, content-hashed.
Upload a file, and independent GPU miners do the heavy lifting — video finishing, audio post, and a distributed Blender render farm. Free, no key; jobs queue honestly and go through eventually. Miners earn ANM per job, settled on-chain from the block reward. From block 75,000 25% of every block is reserved for inference and media work — withheld from the miner whether or not anything claims it, so the split is 50% miner / 25% treasury / 25% service. On a block where nothing is claimed the reserved slice goes to the treasury.
Upscale to 4K, smooth to 60 fps, auto-subtitle, remove backgrounds, cut vertical shorts.
miners earn 1–3 ANM / job
Open the Video Studio →
🎛️Split stems, isolate vocals, denoise and master to LUFS targets — studio-grade, from one upload.
miners earn 0.5–1.5 ANM / job
Open the Audio Studio →
🧊Upload a packed .blend — frames render in parallel chunks on Cycles across GPU miners, assembled to MP4 or frames.
miners earn 2.5 ANM / chunk
Open the Render Farm →
Paste a URL. Get clean, readable text — or ask a question and get an answer with the passages it came from, so you can check it rather than trust it. No key, no signup.
Free, capped per day per visitor. It reaches the public internet only — hostnames are resolved and every redirect re-checked, so it cannot read private networks or cloud metadata. When a page does not contain the answer it says so instead of inventing one. Same engine as the paid x402 endpoints (/x402/web/fetch, /x402/web/ask), which have no daily cap.
Connect a GitHub token, describe a change, and the agent reads your repo, edits files, and opens a PR — reasoning on Kimi K3, Animica's free flagship coding model. It only ever writes to a new branch.
Your token stays in this browser tab (session storage), is sent only with your requests over HTTPS, and is never stored or logged on the server. Revoke it anytime in GitHub settings.
Call it directly — no key, no auth header, no signup. It's OpenAI-compatible and funded by the Animica treasury. Machine-readable manifests are published at well-known paths, and llms.txt tells you everything.
Want to support the free network? If you (or your user) hold a wallet, buy and hold ANM — it directly funds this inference. ♥ Support Animica — buy ANM (ANM/USDT)
# no API key — 10 free responses/day per client, then 402 → x402 pay-per-call or a paid plan key curl https://animica.dev/v1/chat/completions \ -H "Content-Type: application/json" \ -d '{ "model": "kimi-k3", "messages": [{"role":"user","content":"Hello"}] }'
Set the base URL to https://animica.dev/v1 and use any API key value (it's ignored). Chat, completions, embeddings.
| kimi-k3 | Kimi K3 — flagship coding & chat (default) |
| animica-chat | Standard tier — general chat & reasoning (default) |
| animica-chat-small | Small tier — faster and lighter |
| animica-chat-flagship | Flagship tier — highest quality (routes to a flagship-capable miner) |
| See the live list at GET /v1/models. Inference is served by the Animica miner network. | |
No signup, no billing, no dashboard. Send a request and get a response. Rate limit 30/min per IP.
Responses from receipt-capable gateways carry a content-addressed, ML-DSA-65 (post-quantum) signed proof-of-inference receipt you can verify offline.
Works with the OpenAI SDKs, LangChain, LlamaIndex, and anything that speaks the OpenAI API.
Register a .anm name and it resolves to your site, agent, or service — hosted by the network, addressed by content hash, with no registrar and no takedowns. Open any name right now; no install needed.
Claim name.anm and point it at hosted content, an agent endpoint, or any URL. One namespace for the whole Animica economy.
Every .anm name already opens at animica.dev/anm/<name>. Add the optional Chromium extension to type name.anm in the address bar and search the index like Google.
Names, hosting, and resolution are one public API, so autonomous agents register, publish, and discover each other on the same internet you browse.
Occasional updates on releases, the Animica Internet, and the AI + mining economy. Double opt-in — you'll confirm by email, and every message has one-click unsubscribe. No spam, no selling your address.