OpenAI-compatible
Drop-in /v1/chat/completions with SSE streaming. Keep your SDKs and change one base URL.
Route, budget, and audit model traffic across providers. trilinkai gives teams a single OpenAI-compatible endpoint with hard spend controls and a ledger you can reconcile.
OpenAI-compatible · SSE streaming · Team budgets
curl https://api.trilinkai.com/v1/chat/completions \
-H "Authorization: Bearer $TRILINKAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "frontier-reasoning",
"messages": [{"role": "user", "content": "Ship it."}],
"stream": true
}'100+
models behind one API
3
regions planned
10
locales at launch
0
surprise overages
Platform
Drop-in /v1/chat/completions with SSE streaming. Keep your SDKs and change one base URL.
Cost-aware routing, retries, and fallbacks across providers. Failover without changing client code.
Shared team wallets, per-member limits, and a hard stop at the gateway before spend runs away.
Every request maps to a spend log and an immutable ledger entry. Daily reconciliation, no double charges.
Region-aware deployment, English and Chinese at launch, and a path to EU and APAC data residency.
Role-based access, key lifecycle, and append-only audit trails for finance, security, and compliance.
Catalog
Prices, context windows, and availability live next to the route. Preview catalog shown; production models land with Phase 2.
| Model | Context | Route | Status |
|---|---|---|---|
| frontier-reasoning | 200k | Cost + latency | Preview |
| long-context | 1M | Lowest cost | Preview |
| fast-coding | 128k | Lowest latency | Preview |
| open-weights | 64k | Fallback | Preview |
Team controls
Set one shared budget for the team and a per-member ceiling. trilinkai checks the budget at the request edge, then settles from the ledger.
Team budget
$2,400 of $3,000
80% used · hard stop enabled
The foundation is open. The gateway, ledger, and team controls are next.