|
GTokens — one gateway for frontier models
One gateway for frontier models, with published per-token rates by model and token type. Prepaid balance, no subscription, no monthly minimum.
|
82 % |
|
ClaudCli — one API for every model you need
One endpoint for every model, with prepaid tokens and no subscription. Point the client you already use at it — and there is no way to spend past your balance.
|
82 % |
|
public.supply — the supply layer for AI inference
One OpenAI-compatible endpoint across every model provider. Capability-aware routing, per-request budgets, and usage receipts you can audit.
|
81 % |
|
pumdoctrine — don't believe the AI. Check it.
The governing layer under an AI suite: every action receipted, every claim sourced, every limit published. Verify it in your own browser — no install, no trust.
|
81 % |
|
Hookie — Receive, store, route and deliver webhooks
Hookie receives, stores, routes and delivers webhooks: every payload kept whole and searchable, forwarded to your endpoints signed, retried and replayable. F...
|
80 % |
|
Zumik: one API key for every AI model
Zumik is one API key that works with every major AI model. Point your app or coding tool at Zumik instead of OpenAI and it sends each request to whichever mo...
|
80 % |
|
Unify API — One API. 300+ models. Lower costs.
One key and one bill across 300+ models from every major provider, with smart routing and per-request data policies you control.
|
80 % |
|
Switchyard — one AI API for every application you run
Switchyard sits between your applications and every AI provider. One API and one key per app, with routing, budgets, failover and cost accounting handled cen...
|
80 % |
|
BlazeRail — One API for every model
One API for every model, and your OpenAI code already works with it. Change the base URL and keep everything else. Each request goes to the cheapest provider...
|
79 % |
|
VIXTUR — Say it. See it live.
VIXTUR builds real software from a description: a database, accounts for your own users, permissions enforced on the server, and a live address.
|
79 % |
|
Saible: what your AI costs, and what it did
Connect the AI providers you already pay for. Thirty days of spend, contracts and evidence readiness, then the enforcement point for what a bill cannot tell...
|
79 % |
|
Aaly — multi-tenant backend for AI-built apps
Your coding agent defines the data model. Aaly serves the production REST API with sign-up, auth, multi-tenant isolation and validation already applied.
|
79 % |
|
Tīrtha: we stand behind the answer
Most AI gateways forward your request. Tīrtha stands behind the answer: we check it against your tests and your schema, move up to a stronger model only when...
|
79 % |
|
ModelAssay — Provable LLM savings
ModelAssay cuts your LLM bill without touching quality — a drop-in proxy that routes, caches, and batches, proves the savings on live traffic, and only bills...
|
79 % |
|
lowcostllm: frontier models at half price
Claude and GPT, through the official APIs, at up to 50% off list. Not a Chinese model, not a proxy to one. Change one line and keep everything else.
|
79 % |
|
InferClear — OpenAI-Compatible AI Inference API Gateway
One OpenAI-compatible endpoint for every frontier AI model, with a per-request cost ledger, hard budgets and audit logs built for engineering teams.
|
79 % |
|
Nozzle · One key in front of every way a model can be served
An OpenAI- and Anthropic-compatible API: one key calls every model, served from our provider accounts, shared open-weight GPUs, your own provider keys or a G...
|
79 % |
|
OpenVet — supply-chain audits
|
78 % |
|
Hiwayz — the governed AI model router
One OpenAI-compatible endpoint across every frontier model, with policy, spend caps and a tamper-evident audit trail enforced before the request leaves your...
|
78 % |
|
Sortance — one API that decides who gets in
Verify the address, enrich the person, score fit against your rubric and apply the signup rule you wrote in plain English. One call, one decision, with reasons.
|
78 % |
Detected social media networks