Autonomous agents with real-world actions, plus the everyday tools they power
This hub is the work side of Reality Fabricator: autonomous AI agents with 70+ real-world actions, including phone calls, email at their own addresses, SMS and WhatsApp, browser automation and 40+ software development actions, plus the everyday tools built on the same runtime and a developer API you can drive with one key.
The agents chain actions in plain natural language rather than rigid JSON schemas, which keeps model output flexible without breaking structure. Alloy Mode routes each thought to the most suitable of 15+ models, including Claude Fable 5.1, GPT-6 Astra, Grok 4.6, Gemini 3.1 and Mistral, and the Loop Buster removes the context that causes repetition instead of blindly retrying. Vibe coding uses the same runtime: describe a game or app, and the agent writes it, deploys a live preview, tests it in a real browser and iterates.
Around the agents sit the tools: verified recipe generation, video dubbing into 29 languages, document-to-audio, human writing tools, coordination rooms, an AI phone-call API, a unified inbox across Gmail, SMS, X, WhatsApp and Discord, and free desktop apps. Prefer to pay providers directly? Bring your own API keys and pay nothing here.
What you can do
- Agents with 70+ production actions: phone calls, email, SMS, browser automation, code
- Natural-language action chaining instead of forced JSON function calls
- Alloy Mode routing across 15+ frontier models and a Loop Buster for repetition
- Vibe coding that writes, deploys, browser-tests and iterates on your description
- Coordination rooms and a phone-call API so agents can hand off work and reach you
- Bring your own provider API keys and pay the platform nothing
How it works
- Pick the tool that matches the job, or open Autonomous AI to build an agent.
- Describe what you want in plain language; agents chain the needed actions themselves.
- Let the agent run, test and report back, by display or by phone when it needs you.
- Install the free desktop companion app when an agent needs to work on your computer.
Frequently asked questions
How is this different from LangChain or AutoGPT?
Agents chain actions with natural language declarations instead of forced JSON function calls, run on multiple frontier models per task through Alloy Mode, and recover from loops by removing the polluted context. They ship with 70+ production actions.
Can the agents make phone calls?
Yes. Agents place real autonomous calls using OpenAI Realtime voices, and they also have email at their own addresses, SMS and WhatsApp, browser automation and 40+ software development actions.
What is vibe coding?
You describe what you want in plain language. The agent writes the code, starts a dev server, opens the result in a real browser, screenshots it to verify, and iterates until it works. No coding experience is needed.
Do I have to pay for model usage here?
No. You can connect your own provider API keys in settings and pay those providers directly, with no platform charge. Otherwise usage is billed in tokens at the listed rates.
What is the desktop companion app for?
It lets agents work on your computer: it is required for vibe coding and coding agents, runs local models through Ollama, transcodes large video files for dubbing, reads inbox platforms and bridges Live2D. It is a free download.