Projects / TlabRouter
PRIVATE MVP · ACTIVE

TlabRouter

A self-updating AI model discovery and routing platform that searches for providers with free access, verifies the claims, keeps a normalized registry, and routes across models that are actually usable.

DiscoveryVerificationRoutingOpenAI-compatible
discover()
→ candidate providers
→ free quota claims
→ model catalog

verify()
✓ authoritative source
✓ usable endpoint
ROUTE

provider A · ready
provider B · quota
provider C · offline

→ best usable route

The problem

Lists of “free AI APIs” go stale quickly. Providers change quotas, models disappear, signup conditions change, and a search result is not the same thing as a routable endpoint.

TlabRouter treats discovery and verification as separate systems: the web can suggest candidates, but only verified data becomes routable.

Design principles

01
Discovery is not truth

Search creates candidates. Verification decides what enters the registry.

02
Free access has types

Recurring quota, trial credit and genuinely free access are classified separately.

03
Keys stay local

Provider credentials should remain on the user’s machine whenever possible.

04
One useful local endpoint

The router aims to expose an OpenAI-compatible local API where feasible.

STATUS

Private while the core gets boring.

The repository is intentionally private during early MVP development. Once the discovery, verification and routing boundaries are stable enough, the public story can be much cleaner.