Run an AI workforce without giving up your data.
Tangison Agent is a self-hosted AI agent platform. Your prompts, your transcripts, and your tool calls never leave your network. Bring your own OpenRouter key and pick the model. No markups, no lock-in, no silent training on your work.
Built by Tangi Iigonda in Windhoek, Namibia.
Why self-host
What breaks when AI runs outside your infrastructure
01
Your data leaves your walls
Cloud AI services train on your prompts, store your transcripts, and decide when to deprecate the model you depend on.
02
You rent your intelligence
Per-token pricing, rate limits, and surprise bills. The vendor can change terms or pull access at any time.
03
You cannot audit the work
Closed agents hide their reasoning, their tool calls, and their file access. You cannot prove what they did or did not do.
How it works
Four steps from message to result
Send
Message your agent from Telegram, WhatsApp, or the dashboard.
Plan
The agent breaks your request into ordered steps with tool calls.
Execute
Skills run inside your infrastructure, with full audit logs.
Deliver
The agent returns the result and a complete trace of what ran.
Features
Everything you need to run an agent in production
Every capability exists because a real deployment failed without it. No filler, no features you have to disable to feel safe.
Sovereign by default
The agent process, the workspace, and the audit log all live on your hardware. No third party ever sees your prompts.
Bring your own model key
Plug in your OpenRouter API key. Pick the model that fits the job. Tangison never resells credits and never marks up your usage.
Composable skills
Seven built-in skills cover web search, file I/O, code execution, HTTP fetch, messaging, deploys, and MCP connectors. Add your own in TypeScript.
Meet your users where they are
Ship the agent to Telegram, WhatsApp, and the web dashboard in one config. Same brain, every channel, no extra bill.
Open source runtime
The Tangison Agent runtime is open source under MIT. Read every line. Fork it. Run it offline. Contribute back. No black boxes, no telemetry you did not write yourself.
Full operational control
Workspace isolation, per-skill network policies, and a complete task log. You see every input, every output, every side effect.
Use cases
Where Tangison Agent earns its keep
A project by
Tangison · tangison.com
Tangison Agent runs on your hardware, calls your model, and keeps your audit log on a disk you control. The runtime is MIT licensed. Fork it, own it, walk away with everything if you ever decide to leave.
Run your first agent on your own hardware
Tell Tangi what you want to automate, what infrastructure you run, and which model you want to use. Message him directly on WhatsApp and he replies within one business day, Monday to Friday, 09:00 to 17:00 CAT.