Gateway
One standard chat-completions API in front of open-weight models, premium models and your own keys.
OpenUno Core · In design
The open-source AI gateway and model router
Run one gateway in front of every model you use, on your own servers. Routing, budgets, keys and observability, open at the core.
What is inside
Six modules, one deployable engine
Router
Picks a model per request by quality, cost, latency and your rules, with automatic fallback.
Budgets and limits
Spend caps and rate limits per key, user, team or workspace, enforced before the request leaves.
Keys and identity
Issue scoped keys, bring your own provider keys, and connect your identity provider.
Observability
Usage, latency and cost for every request, exportable to the tools you already run.
Tools and connectors
A common interface for tools, data connectors and agents on top of the router.
Developer experience
Set the model to auto. Set a budget. Ship.
Keep the API your code already speaks. Ask the router to choose, give it a preference and a ceiling, and every response tells you which model answered and what it cost.
UnoThis request shape is the design target, not a released API. Field names may change before the first release.
curl https://your-gateway.example/v1/chat/completions \
-H "Authorization: Bearer $OPENUNO_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "auto",
"messages": [{ "role": "user", "content": "Summarise this report" }],
"routing": { "prefer": "cost", "max_usd": 0.01 }
}'Documentation
Documentation arrives with the first Core release. Early-access members get the design notes and preview builds first.
Get notifiedLicence
Core will be released under an open-source licence. The licence is chosen and reviewed before the first public release, and will be stated here.
Help shape the open core.
Early-access developers see the design first and tell us what to build next.