liveuprouter — AI API router and gateway
Find the best AI API routers.
Use them all with one API key.
UPROUTER is an OpenAI- and Claude-compatible gateway: plug in your API keys from 310 routers (eg. OpenRouter, AWS Bedrock, Alibaba Cloud), then route every request with any model, through any vendor — via a single URL with failover aliases, daily spend caps and per-request compute metering you can audit line by line.
- BYOK keys — AES-256-GCM encrypted
- failover aliases, cheapest-first
- daily caps with soft-limit headers
- per-request compute metering
$ curl https://uprouter.online/api/connect/v1/chat/completions \
-H "Authorization: Bearer upr_live_…" \
-d '{"model": "sandbox-echo"'}'
← 200 OK · 412ms · 0.0002 compute · routed via sandbox-echo
audit log: tokens in/out, latency, cost — per request
{
"id": "chatcmpl_9f2…",
"model": "sandbox-echo",
"routed_via": "sandbox-echo",
"usage": { "compute": 0.0002, "tokens": "42/310" }
}refer a developer+50 compute
they get 10 — you get 50
310 routers & providers · 490 models tracked · $4.3k free credits
Every entry answers to the community
The index is expansive; the community keeps it honest. Votes, reviews and verified corrections feed straight back into rankings and data confidence.
- Vote
Upvote routers that deliver, downvote the ones that don’t — the score re-ranks the index.
- Review
Evidence-based reviews with verification tags. Ratings, not vibes.
- Verify
Corrections that survive review earn compute — and flip the entry to verified.
Anyone can edit this index
Every router page has an edit button — propose corrections to descriptions, prices and free tiers. Editors review each proposal before it goes live, and accepted edits earn compute.
- Propose
Open a router page, hit the edit button, and detail your proposed corrections.
- Review
Editors rigorously verify every proposed correction against primary sources.
- Live
Accepted edits are published instantly, and you earn compute for your work.
$ 22 entries match — showing 1–22 · sort: Free credits ↓
- Free relayup385msupdated 14h agoapi.bluesminds.comsource · api.bluesminds.comoverview
BluesMinds is a primarily-paid AI relay reselling access to Claude and GPT at significant discounts. It offers a large free credit bonus ($500) for accounts linked to aged GitHub profiles, and provides free daily calls for specialized models like GPT-4.1.
free quotaUp to ~$500 in credits for aged GitHub accounts per community reports.
free models— none trackedsee api.bluesminds.com’s free quota / risk review → - Free relayup997msupdated 13h agoagentrouter.orgsource · agent-router.org/pricingoverview
Agent Router is an open-source, public-welfare AI API aggregation platform designed to unify access to agents and MCP tools. It is 'forever free' for both users and developers, providing a discovery registry for skills (ClawHub) and a live A2A directory.
free quotaOffers $100-$200 free credit bonus for new accounts to bootstrap agent usage.
free modelsClaude (full Claude Code family)DeepSeekZhipu GLMgpt-5see agentrouter.org’s free quota / risk review → - Free relayup204msupdated 14h agoapi.gemai.ccsource · api.gemai.ccoverview
Hajimi API (api.gemai.cc) is a Chinese third-party relay aggregation platform. It specializes in routing requests to Claude, Gemini, and GPT models through a unified OpenAI-compatible endpoint, primarily catering to users requiring simplified access to international models.
free quotaNew users get 100 units of balance by default (community figure ~¥500 (≈$74.39)), QQ-email signup only; there is a daily check-in farming crowd; 'limited perk' zero-priced models can be delisted anytime.
free modelsGemini seriesGPT seriesclaude-sonnet-4-6claude-opus-4-6claude-opus-4-7claude-opus-4-8see api.gemai.cc’s free quota / risk review → - Free relayup1320msupdated 13h agoanyrouter.topsource · anyrouter.topoverview
AnyRouter (anyrouter.top) is a public-welfare AI API relay primarily serving the mainland Chinese developer community. It specializes in routing Claude Code and other frontier models for free, supported by community sign-in bonuses and referral credits.
free quotaGrants ~$50-$100 in free credits on sign-up based on community reports.
free modelsClaude 4 OpusClaude 4 Sonnetsee anyrouter.top’s free quota / risk review → - Free relayup284msupdated 14h agoapi.feifeimiao.topsource · api.feifeimiao.topoverview
Feifeimiao (api.feifeimiao.top) is a Chinese third-party relay service offering an OpenAI-compatible API gateway. Built on the New-API framework, it aggregates various models and focuses on providing access to tools like Codex and Claude Code for developers in the region.
free quota$10 trial credit on signup; promotional $20 weekly-limited cards for new and existing users (weekly expiry). The core is paid subscriptions (Small ¥60 (≈$8.93)/30d ≈ $20/day, Medium ¥150 (≈$22.32)/30d ≈ $60/day).
free modelsgpt-image-2claude-opus-4-7gpt-5.5see api.feifeimiao.top’s free quota / risk review → - Free relayup854msupdated 14h agomonkeyapi.apimonkey.onlinesource · monkeyapi.apimonkey.onlineoverview
MonkeyAPI (monkeyapi.apimonkey.online) is an anonymous individual relay service built on the New API framework. In late 2026, the service operator reported penalties for violations and announced plans to migrate, significantly reducing quotas and referral benefits.
free quotaPreviously offered registration credit and check-in/referral bonuses, but notices state it is 'cutting referral and registration quotas for cost control' and will adjust after migrating; no stable public free flagship quota.
free models— none trackedsee monkeyapi.apimonkey.online’s free quota / risk review → - Free relayup106msupdated 14h agofreetheai.xyzsource · freetheai.xyzoverview
FreeTheAi is a free OpenAI-compatible API gateway aggregating 80+ models behind a single base URL. It operates via Discord-based key signup and daily check-ins, emphasizing persistent free access for developers and roleplay users without credit card requirements.
free quotaFully free, no card; daily /checkin (key validation + human challenge) required before calls. Default cap 250 successful calls/day (role-based bumps) ~ $7.5/mo equivalent at ~1K tokens/call, $1/1M tokens.
free models— none trackedsee freetheai.xyz’s free quota / risk review → - Free relayup2333msupdated 13h agoapi.iamhc.cnsource · api.iamhc.cnoverview
IAMHC (Xinjiang Huancheng Wangan) is a public-welfare AI API gateway providing unified access to over 199 mainstream models. It features a rate-limited free tier and emphasizes compliance with built-in moderation for developers in the China region.
free quotaFree tier: small-parameter models only, ~1500 calls per 5h (some sources say ~800), unguaranteed and test-only; paid pub/VIP/SVIP tiers. Rough ≈$5/mo equivalent; reliability low.
free modelsFree tier = small-parameter test models (specific names not individually disclosed)Paid tiers cover 199+ mainstream models including GPT/Claude/Geminisee api.iamhc.cn’s free quota / risk review → - Free relayup224msupdated 14h agovsllm.comsource · vsllm.comoverview
VSLLM (Weiyun Model Open Platform) is a commercial AI API relay gateway that aggregates multiple large models. It is OpenAI-compatible and features a daily auto-refreshing free quota for new sign-ups. Claims to offer highly competitive, transparent pricing based on actual token usage.
free quotaFree subscription auto-refreshes daily without claiming: ~20 API calls + ~100K tokens/day (non-cumulative, reset next day) ~ $3/mo at $1/1M tokens; also GitHub-star and daily check-in reward activities.
free modelsgemini-3.1-pro-request-antigravitymimo-v2.5-tts-voicedesignglm-5.3-anthropicclaude-opus-4-5-20251101gemini-3.5-flash-antigravitycodex-auto-review+33see vsllm.com’s free quota / risk review → - Free relayup349msupdated 13h agochatanywhere.techsource · chatanywhere.tech/pricingoverview
ChatAnywhere is a widely used free LLM relay. Users bind a GitHub account to get a free API key and access GPT-4o, Claude, and Gemini models through an OpenAI-compatible interface, optimized for access from mainland China.
free quotaGitHub-bound free key: 10,000 points/day with per-model caps (GPT-5/4o/4.1 5/day, DeepSeek 30/day, mini models 100/day, ~200 req/day total) ≈ $2/mo; personal non-commercial use only.
free modelsgeminigrokQwenKimiminimaxclaude+7see chatanywhere.tech’s free quota / risk review → - Free relayup359msupdated 13h agoapi.chatanywhere.orgsource · chatanywhere.techoverview
ChatAnywhere (GPT_API_free) is a widely used AI API relay providing unified access to global models including OpenAI, Claude, Gemini, and DeepSeek. It offers a GitHub-authorized free tier and low-cost paid access with endpoints optimized for users in mainland China and overseas.
free quotaFree keys, personal non-commercial only: ~200 req/day per IP&key (GPT-5/4o/4.1 5/day, DeepSeek 30/day, small models 200/day) ≈ $2/mo equivalent at typical token volumes.
free modelsGPT-5 series (5/day)DeepSeek (30/day)several small modelsGPT-4ogpt-4.1see api.chatanywhere.org’s free quota / risk review → - Free relaydown427msupdated 14h agoapi.apiplus.cloudsource · api.apiplus.cloudoverview
APIPlus is a third-party AI API relay built on the New-API framework, reselling access to Claude, GPT, and Gemini. It offers a small registration credit and an intermittently available 'free channel' with deep discounts, primarily targeting the low-cost reseller market.
free quotaGrants ~$1 registration credit for new users.
free modelsGPT family (free channel at 0.01x, not free)see api.apiplus.cloud’s free quota / risk review → - Free relayup1505msupdated 13h ago9router.xunleiyk.comsource · 9router.comoverview
9Router is a specialized AI API router designed for smart fallback and cost optimization, allowing users to pool their own subscription keys. It emphasizes 'unlimited free AI coding' by routing through free models or maximizing existing subscriptions via a local proxy or VPS.
free quota9Router is free open-source self-hosted router software (never charges); 9router.xunleiyk.com is a hosted instance ('AI Infrastructure Management' panel, auth-gated). No credits/billing tiers apply.
free models— none trackedsee 9router.xunleiyk.com’s free quota / risk review → - Free relayunknownupdated 3d agofreebuff.comsource · freebuff.comoverview
Freebuff is a specialized AI API relay often associated with developer automation and coding tools. It provides programmatic access to high-end models for free or via community-shared keys, primarily aimed at script-based integration and tool builders.
free quotaNo public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.
free modelsgpt-5.6-lunaDeepSeek V4 FlashSolar Pro 4deepseek-v4-prosee freebuff.com’s free quota / risk review → - Free relayunknownupdated 3d agohelyxai.spacesource · helyxai.spaceoverview
Helyx AI is a multi-model gateway providing access to 50+ frontier models like Claude 5 and GPT-4o. It features a unique 'three-stage waterfall' quota system including a large daily free claim, a permanent pool earned via referrals/reviews, and a paid credit balance.
free quotaNo public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.
free modelsClaude Haiku 4.5deepseek-v4-flash-0731Qwen3 32BGemma 4 31Bgemini-3.1-flash-liteMiniMax-M3+4see helyxai.space’s free quota / risk review → - Free relayup273msupdated 13h agoapi.llm7.iosource · llm7.iooverview
LLM7.io is an open-source AI API gateway providing unified access to leading models like GPT and DeepSeek. It stands out for its unique anonymous access tier, allowing developers to test models without registration, and a registration-based free tier with higher limits.
free quotaFree relay: ~30 RPM anonymous, rising to ~120 RPM with free token from token.llm7.io (20/min, 100/h sub-caps); base URL api.llm7.io/v1 with api_key 'unused'. No token quota published → no USD estimate.
free modelsDeepSeek V4 Flashkimi-k2.6kimi-k3minimax-m2.7gpt-oss:20bcodestral-latest+4see api.llm7.io’s free quota / risk review → - Free relayup644msupdated 14h agonewapi.makelove.cloudsource · newapi.makelove.cloudoverview
MakeLove NewAPI (Nianhua API) is an anonymously operated Chinese API relay on the New-API panel that aggregates various models with paid top-up packages as the primary business model. Details were checked against the provider's own page at newapi.makelove.cloud.
free quotaFree model entries remain (~13 per /api/pricing); affiliate posts advertise '$10 signup + daily check-in' but site announced removing new-user free credit; free models are the reliable free tier.
free modelsgemini-3.5-flashGemini 3.1 ProGrok 4.3see newapi.makelove.cloud’s free quota / risk review → - Free relayunknownupdated 3d agobynara.idsource · studentoffers.co/offer/nararouteroverview
NaraRouter (bynara.id) is a regional AI relay providing generous free token allowances (up to 7M tokens/day) for students and hobbyists. It supports an OpenAI-compatible API and Google login, resetting quotas daily at 07:00 WIB.
free quotaNo public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.
free modelsDeepSeek V4 Flashglm-5.3-flashsee bynara.id’s free quota / risk review → - Free relayup223msupdated 14h agollm.onallways.topsource · llm.onallways.topoverview
OnAllways (llm.onallways.top) is a 'subscription-to-API' relay that pools various AI subscription accounts (Claude, GPT) and redistributes them as API keys. It is anonymously run and targets users looking for low-cost access to restricted models.
free quotaUnknown: the site's subtitle is 'Subscription to API Conversion Platform', with payment/affiliate/risk-control enabled and traffic funneled to third-party card shop catfk.com; no clear public persistent free quota is visible.
free modelsClaude seriesCodex/GPT seriesGemini series (forwarded via pooled subscription accounts, dubious authenticity/stability)see llm.onallways.top’s free quota / risk review → - Free relayup274msupdated 14h agofree.supxh.xinsource · free.supxh.xinoverview
Shawn AI (肖恩AI) is a community-driven API relay station aimed at roleplay and tavern users. It aggregates hundreds of models including Claude 4.7 and GPT-5.5, offering free daily points and signing bonuses. It features a transparent point-deduction system and recovery for failed generations.
free quotaUses a 'points' system (5000 on signup, 1000-3000 daily check-in) which does not have a fixed USD conversion rate.
free modelsclaude-opus-4-5/4-6gpt-5.2/5.4/5.5qwen3.7-max-tGemini 2.5 Prodeepseek-v4-prominimax-m2.7see free.supxh.xin’s free quota / risk review → - Free relaydown224msupdated 14h agoapi.xstx.infosource · freeairouter.com/s/api.xstx.infooverview
Xingtu AI (api.xstx.info) is a third-party LLM aggregation relay that was previously marketed for free Claude and Codex access. Community reports and service status trackers indicate the service has been down for several months and may have effectively 'soft-runaway', transitioning to a new platform named Reclaude.
free quotaNo public free-tier information found (researched 2026-08-28). Source note: Previously attracted users with 'free Claude/Codex'; its companion chat client chat.xstx.info had a shared daily 100-CNY quota; the API side uses CNY pricing with Stripe top-up (~$8/unit). Current free availability is doubtf
free models(claimed) Claude family(claimed) Codex/GPT familysee api.xstx.info’s free quota / risk review → - Free relayunknownupdated 3d agog4f.spacesource · g4f.spaceoverview
g4f.space is an AI API router providing access to multiple frontier models including Gemini. It emphasizes unrestricted, open-source access but now requires proof-of-work credits for anonymous use to maintain service quality.
free quotaNo public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.
free modelsGemini 3.8 Flashgemini-3.7-flashsee g4f.space’s free quota / risk review →
UPROUTER.ONLINE is an independent, developer-facing directory of AI API providers. It tracks model aggregators, official first-party APIs, free relays, subscription routers and cloud platforms in one index. Inclusion is an editorial decision made on the merits of the data — it is never sold, and no listing position can be bought.
We do the normalization once so you can compare on facts instead of marketing: prices are expressed in USD per 1M tokens, free tiers are valued at each provider’s own pay-as-you-go rates, and reliability comes from our own best-effort uptime probes rather than self-reported status pages. Every entry carries source URLs, a verification date, a data-confidence flag and a change history.
The directory is informational only and is not an endorsement of any provider. Prices and free-tier values change frequently — always confirm current terms with the provider before you purchase. The full methodology, independence and non-endorsement policies are documented in detail, and you can reach the editors via /contact for corrections, takedowns or partnership questions.
$ free AI API key — a practical guide
Everything below is written for the person who just typed free AI API key into a search box: what the key is, how to get one, what “free” really costs, and how to use more than one without managing six dashboards. It is the prose companion to the live index above, which tracks the same providers with current numbers.
What is a free AI API key?
An AI API key is the credential your program uses to call a language model. Think of it as a password for a specific service: it identifies you, authorizes the request, and is what the provider meters and bills. A free AI API key is the same object, except the usage behind it is covered by the provider rather than your credit card — at least for a while.
Under the hood, the flow is always the same. You create a key in a provider’s dashboard, you send it with each request, and the provider runs your prompt on a model and returns tokens. The free part is a policy layer on top of that plumbing: a monthly credit budget, a per-minute and per-day rate cap, or a restricted list of models marked free. None of those policies change the fact that a free key and a paid key are the same shape of string pointed at the same kind of endpoint.
The reason the distinction matters in practice is that a free key is usually metered against a budget you did not choose to spend. When the budget empties, your request stops working until you reset the period, top up, or switch models. A paid key is metered against money you explicitly put there. Everything else — the header, the JSON request, the streaming response — is identical, which is exactly why a single gateway can sit in front of both.
How to get a free AI API key in 2026
Getting a free key takes minutes. The steps are the same at nearly every provider, so you can pick whichever free tier fits your use case and follow the pattern:
- Choose a provider. If you want the easiest first experience, Google AI Studio or Groq are the most forgiving. If you want the widest choice of models behind one key, OpenRouter is hard to beat. The directory above lists each option with its current free-credit value so you are not guessing.
- Create an account. An email address is enough for most free tiers. A handful ask for a card on file for abuse prevention but will not charge you until you move to a paid plan — read that note before you enter it.
- Generate the key. In the dashboard’s “API keys” or “credentials” page, click create. The key is shown once; copy it to a secrets manager or environment variable immediately. Do not paste it into client-side code or a public repo.
- Make a test call. Send a tiny completion request to confirm the key works and the free tier is active. If the response includes a usage or rate-limit header, note the numbers — those are the real limits you are working within.
- Decide where to keep it. Either point your application straight at the provider, or drop the key into a gateway like Uprouter Connect so you can add more keys, set spend caps, and fail over when a free tier throttles you.
One nuance worth internalizing: the value of a free key is not that it exists, it is that it is portable. Because every free key talks the same OpenAI- or Claude-compatible protocol, the code you write against one key will run against another with only the base URL and the credential changed. That portability is what makes it sensible to hold several free keys and route among them.
Free key vs. free credits vs. free tier
People say “free AI API key” and “free credits” interchangeably, but they are three different things, and confusing them is how a project quietly stops working at exactly the wrong moment.
- The key is the credential. It has a value only in that it grants access to whatever budget is attached to your account.
- Free credits are a dollar amount of usage the provider has fronted for you — a signup grant or a trial balance. They are one-time in most cases: spend them and they are gone until the provider replenishes them.
- A free tier is a standing, rate-limited plan you can use month after month. It is usually the “always available” part of a free key, with credits layered on top for a burst.
When you evaluate a provider for a free key, ask three questions. First, how many credits do I get on signup, and do they ever come back? Second, what is the standing free tier — which models, and what are the per-minute and per-day caps? Third, is there a hard wall — a day when the key requires payment to keep working? The directory above values the free tier at each provider’s own pay-as-you-go rates so the “how much is this actually worth” question is answered in one comparable number.
Safety and terms: what “free” actually means
The three things a free key’s terms almost always get wrong for the user are privacy, durability, and data use. Read for all three before you build on top of a free tier.
Data and training
Many free tiers may use your prompts and completions to improve models; paid plans usually may not, or let you opt out. If your traffic contains anything you would not want a model vendor to see, this single clause can disqualify a free key for production regardless of how generous the credits are.
Durability
A free tier is a marketing surface, not a service level. Limits can be reduced, models can rotate, and free previews can end. Treat any free key as best-effort and keep at least one alternative key warm so an upstream change is a config swap, not an outage.
Responsible use
Rate limits exist for the provider’s cost and your neighbors’ latency. Hitting them aggressively, or routing shared keys across many users, is the fastest way to get a free account suspended. Meter your own calls, cache what you can, and keep a per-user or per-project cap so one caller cannot burn the shared budget.
The risk tier on each entry in the directory above is our editorial read of exactly these questions — established operator versus young relay, clear terms versus vague ones, stable history versus a history of limit changes. It is one signal among many, and it is always paired with the source so you can verify it yourself.
The most useful free AI API keys right now
A shortlist to start from, grouped by what the free tier is best for. Free tiers change often, so treat the notes as the shape of the offer, not a spec — the live index above has the current credit values, rate-limit observations, and risk tier for each.
| Provider | Models | Free offer | Notes |
|---|---|---|---|
| Google AI Studio | Gemini API | Free tier with rate limits | Standard models are free to key with no card; heavy or pro models move to paid. Good first key for chat and embeddings. |
| Groq | Llama, Gemma, and more | Free API key, fast inference | One of the lowest-latency free keys around; per-minute and per-day rate limits apply. |
| OpenRouter | Hundreds of models | Free :free model routes | A router with free variants of many models marked :free, plus a credit system for paid ones. |
| Mistral AI | Mistral, Codestral | Free La Plateforme tier | Free-tier keys with rate limits; a paid tier unlocks higher throughput. |
| Hugging Face | Open models | Free inference providers | Free access to a range of open models through hosted inference, subject to per-day quotas. |
| Cohere | Command, Embed | Free trial key | A trial key with a usage budget for text generation, embeddings, and reranking. |
| Cerebras | Llama, Qwen | Free fast inference | Very high token rates on a small set of open models, free while in preview. |
verify current terms with the provider before relying on a free tier in production.
Using one gateway across many free keys
The natural next step after collecting a few free keys is to stop calling each provider by its own URL. Because the keys all speak the same protocol, you can point them at a single gateway and drive them like one API. That is the job of Uprouter Connect: it takes your OpenAI- and Claude-compatible calls, routes them to whichever provider you name, and gives you one base URL and one auth scheme in exchange.
The payoff is threefold. Failover: when a free tier rate-limits you mid-request, the gateway can try the next provider on an alias instead of returning an error. Metering: every call is logged with tokens in, tokens out, latency, and compute cost, so you can see exactly how hard each free key is working. And caps: you can set a daily compute ceiling so a runaway loop cannot drain a shared budget before you notice. For anyone managing more than one free AI API key, that single point of control is usually worth the hop.
Frequently asked questions
What is a free AI API key?
A free AI API key is a credential that lets your code call a large-language-model or AI service without paying, usually within a free tier or a set of free models. You sign up with an email, the provider issues a key, and you send it with each request (typically as an Authorization header) to hit an endpoint such as /chat/completions. "Free" almost always means "free up to a limit" — a monthly credit allowance, per-minute and per-day rate caps, or a subset of free models — not unlimited usage forever.
How do I get a free AI API key in 2026?
Pick a provider that offers a free tier (Google AI Studio, Groq, Mistral, OpenRouter, Hugging Face, Cohere, Cerebras are common starting points), create an account, open the API-keys console in the dashboard, and generate a key. Copy it into your client or into a gateway. Most free keys require no credit card, though some providers ask for one to prevent abuse and refund nothing until you top up.
Are free AI API keys really free, or is there a catch?
They are free to use within the stated limits, and the catch is the limit — not a hidden fee. Typical constraints: a monthly or one-time credit budget, rate limits (requests per minute, tokens per day), a restricted set of models, and data-handling rules that differ from paid plans. The important catches to read before you rely on a free key are (1) whether free-tier data may be used for training, (2) what happens to your key when the trial budget is exhausted, and (3) whether the free tier is guaranteed or best-effort.
What is the difference between a free key, free credits, and a free tier?
A free key is the credential itself. Free credits are a dollar amount of usage the provider grants so you can run paid models at no cost until the budget runs out. A free tier is an ongoing, rate-limited free plan you can keep using. A provider may give you all three at once: a free key, a one-time credit grant, and a persistent free tier for cheaper models. Knowing which of the three you are using tells you what happens the moment the number hits zero.
Which providers have the most useful free AI API keys?
For a first key, Google AI Studio (Gemini) and Groq are the friendliest: no card for standard models and fast responses. For breadth, OpenRouter gives you one key across hundreds of models with explicit :free variants. For open-weight models at speed, Cerebras and Hugging Face are strong. The right choice depends on whether you value raw speed, model variety, or a large credit budget — the directory on this page ranks all of these by free-credit value, risk, and live status so you can compare on facts.
Is it safe to use a free AI API key for a production app?
Use a free key for prototypes, demos, and light internal tooling, but treat it as best-effort: free tiers change limits, can be throttled, and usually sit behind paid plans for production SLAs and higher throughput. Before you ship something user-facing, read the provider’s terms on data retention and training use, confirm your rate limits cover real traffic, and have a paid or gateway-based fallback. If you are juggling several free keys, route them through a single gateway so you can switch providers and meter usage in one place.
What is Uprouter Connect and how does it help with free keys?
UPROUTER CONNECT is an OpenAI- and Claude-compatible gateway. You bring your keys (the free ones from the directory, your paid ones, or a mix), and every request goes through one URL with one auth scheme. That means failover across providers when a free tier rate-limits you, per-request compute metering so you can see exactly what each call costs, and daily spend caps. The directory above maps which providers work through Connect, their free-credit values, and their risk profile, so you can start with the best free key and graduate to paid without rewriting your integration.
Start with a free key, stay on one gateway
Claim welcome compute, add the free keys from this page, and make a real gateway call in under a minute. Every provider in the index above is marked for which ones route through Connect, so you can start free and scale to paid without changing a line of application code.
Informational only — not an endorsement. Verify pricing with the provider. ToS compliance is your responsibility.
Guides on AI API routers, free tiers, pricing and reliability — written by the Uprouter editorial team, data-true and linked to the live directory.
read the blog →