AI API Router Directory

LongCat is Meituan's large model API platform (longcat.chat/platform), offering OpenAI/Anthropic-compatible endpoints. During its public beta phase in 2026, it provides generous daily free token allowances for models including LongCat-Flash and LongCat-2.0, with a context window of 131K tokens.

Free during public beta: initial ~5M tokens/day (per API changelog), activity can raise it up to ~120M/day; ~$150/mo equivalent at blended $1/1M tokens. Phone signup, key in ~3 min; paid top-ups not yet supported.

Pollinations.ai is a Berlin-based open-source platform providing a signup-free, OpenAI-compatible API for text, image, and audio generation. It aggregates multiple open-weights models (DeepSeek, Qwen, Mistral) and offers them for free to foster creative AI development.

Basic use needs no signup and no API key; new Pollen credit system gives ~1.5 Pollen/week free, with some models/higher limits requiring an account or paid tier. Pollen has no published USD rate - not quantifiable.

Agnes AI, developed by Singapore-based Sapiens AI, is an omni-modal AI platform offering indefinitely free access to core text, image, and video models. It supports OpenAI-compatible API keys and tiered 'Token Plans' for production-level throughput.

Advertised as indefinitely free core-model API with rate/volume caps; signup probe hints ~$1 starter credit (counted). Free tier reportedly sufficient for individual developers and demos.

DeepSeek's official API serves the current V4.1-Flash and V4-Pro models with 1M-token context and 384K max output, from $0.15/$0.60 per 1M in/out off-peak (Flash) — peak hours bill at 2x, cached input drops to $0.003/M. One key exposes both an OpenAI-compatible and an Anthropic-compatible endpoint, and sign-up grants a free token balance that is deducted before any top-up.

Free 'granted balance' of ~5M tokens at sign-up (deducted before top-ups per official billing rules); $0.70 is an estimate at V4-Flash off-peak rates.

Aion Labs provides specialized fine-tuned models for immersive roleplay and storytelling. Its API is OpenAI-compatible and features a long-term daily free allowance for developers, with higher tiers available upon account top-up.

Daily allowance of ~20K tokens equates to ~$0.50/mo in nominal value.

Hugging Face Inference Providers routes one API key across 200+ models from multiple inference providers at the providers' own rates — no markup. Every user gets monthly credits ($0.10 on the Free plan; $2.00 PRO; $2.00/seat Team & Enterprise) applied automatically before pay-as-you-go; a BYOK mode routes through your own provider keys instead.

Official docs: $0.10/month Inference-Providers credits for Free users (PRO $2.00/month, Team/Enterprise $2.00/seat), credited monthly and applied before pay-as-you-go (confirmed 2026-09-16).

Adobe Firefly is Adobe's generative image, video and audio product, sold through Firefly plans that bundle monthly generative credits and available to developers as the Firefly API inside Firefly Services. Firefly Standard is US$9.99/month with 2,000 generative credits, Pro is US$19.99/month with 4,000 credits, Pro Plus is regularly US$49.99/month (promotional US$34.97 for the first year) with 10,000 credits, and Premium is regularly US$199.99/month (promotional US$139.91) with 50,000 credits. Adobe publishes no numeric free-tier credit: free users get a limited number of free daily generations across Firefly, Express, Stock, Photoshop on the web, Acrobat and Adobe mobile apps on a curated set of models, with the number of generations, features and models subject to change.

No numeric free credit is stated on official pages: Adobe describes free access only as a limited number of free daily generations on a curated set of models, with counts subject to change; the smallest paid grant is 2,000 generative credits per month at US$9.99/month (adobe.com plans page and Adobe generative credits FAQ, checked 2026-09-16).

Arena.ai (formerly LMArena / Chatbot Arena, lmarena.ai) is a free, anonymous multi-model chat + blind A/B voting service that powers the public LLM/image/code leaderboards. It rotates a large lineup of frontier and open models. This route is an unofficial web-cookie integration (arena-auth cookie, optional Cloudflare clearance/reCAPTCHA tokens); the site also supports anonymous sessions but automated access is rate-limited and subject to bot checks.

A free tier is available; the exact credit value is not quantified.

Auggie is Augment Code's free, terminal-based agentic coding CLI (npm i -g @augmentcode/auggie) that runs context-aware agents locally with interactive and headless modes. It is multi-model (Claude, GPT, or bring-your-own-key); the broader Augment 'Cosmos' platform is the company's enterprise offering.

Free CLI tier; enterprise/Cosmos pricing not publicly quantified on the fetched pages.

Chipotle 'Pepper' is the free, anonymous customer-support chatbot on amelia.chipotle.com, powered by IPsoft Amelia. It requires no login (anonymous sessions) and answers only Chipotle-related questions; off-topic prompts are refused.

Free anonymous support chatbot; no credit system.

Conol (conol.ai) — knowledge-base product with background research/notes agents. Web app forwards to OpenAI models via its model picker (e.g. GPT-6 Astra). Unofficial/experimental.

A free tier is available; the exact credit value is not quantified.

no-auth access — no key or sign-in required. Authentication is owned by the official Devin CLI in its isolated bridge volume. This entry is an access surface/agent — it routes to third-party models rather than hosting its own.

A free tier is available; the exact credit value is not quantified.

Dola (dola.com) — ByteDance's international AI chat assistant (international Doubao) serving the Doubao Seed model family. Unofficial browser/cookie integration.

A free tier is available; the exact credit value is not quantified.

Duck.ai is DuckDuckGo's free, private AI chat (also surfaced at duckduckgo.com/duckchat). It requires no account and offers chat, voice, image, and search across models including OpenAI GPT (e.g., '5.6 Luna') and Claude.

Free anonymous chat; an 'Upgrade'/Pro option exists but pricing not quantified on fetched pages.

Free.ai is an AI tool aggregator offering 477+ tools across chat, image, and video. It provides 30,000 free tokens daily for registered users and supports premium models like GPT-6 and Claude via paid plans. The platform features an in-browser IDE for deploying AI applications.

No public free-tier information found in the source reference; treat free-tier availability as unverified rather than confirmed absent.

The Gemini app is free with a personal Google Account at 'standard limits'; Google's help centre states Gemini Apps use compute-based usage limits that refresh every 5 hours until a weekly limit is reached. Free accounts get a 32k-token context window and access to Gemini Flash-Lite, Flash and Pro models, while paid tiers raise limits (Google AI Plus 2x at $4.99/mo, AI Pro 4x at $19.99/mo, AI Ultra 5x or 20x at $99.99 and $199.99/mo) and lift the context window to 1M tokens on Pro and Ultra.

No monetary credit accompanies the free Gemini tier; Google states free access has 'standard limits' with a 32k-token context window, while paid Google AI plans are $4.99/mo (Plus), $19.99/mo (Pro) and from $99.99/mo (Ultra, 5x and 20x variants).

Grok Build is xAI's terminal coding agent, installed from x.ai/cli and documented at docs.x.ai/build, running on grok-4.6. The official CLI page states the tool is available to try for free and the client is open source, while xAI's developer pricing page meters every request at API token rates: grok-4.6 costs $2.00 per million input tokens, $0.50 per million cached input tokens and $6.00 per million output tokens for prompts under 200k tokens, rising to $4.00 / $1.00 / $12.00 once a request reaches the long-context threshold.

No numeric free allowance is stated - x.ai/cli says only that Grok Build is available to try for free (checked 2026-09-16).

Microsoft Copilot (copilot.microsoft.com) — free consumer Copilot web app with GPT-class chat, web grounding and image generation. Unofficial browser/cookie integration.

A free tier is available; the exact credit value is not quantified.

Microsoft Designer (designer.microsoft.com) is Microsoft's free AI design/image-generation web app (DALL-E based) for personal Microsoft accounts. Free users get a limited number of fast generations (boosts) - historically 15/day, with later region/account-specific limits reported (~60/month). This route is an unofficial, reverse-engineered web integration that captures the short-lived Bearer access_token from the DallE.ashx GetDallEImagesCogSci request; tokens expire quickly and must be recaptured.

Free personal-account boosts only (no credit currency documented).

Meta AI web (meta.ai) is Meta's official free consumer assistant, powered by the Muse Spark family (hosted multimodal reasoning models; current web model is Muse Spark 1.3 per Meta AI developer posts). Free to use with a Meta account, no consumer subscription. This route is an unofficial web integration using the ecto_1_sess cookie plus a short-lived ecto1 WebSocket auth token captured from the meta.ai clippy endpoint.

A free tier is available; the exact credit value is not quantified.

no-auth access — no key or sign-in required. No API key required — uses OpenCode's public free endpoint. This entry is an access surface/agent — it routes to third-party models rather than hosting its own.

A free tier is available; the exact credit value is not quantified.

Qwen Studio is the consumer web app and desktop/mobile client for the Qwen model family, served at qwen.ai and app.chat.qwen.ai. The official product page states that Qwen Studio is free to use and open to all, and the same page lists Thinking, Search, Deep Research, Web Dev, Artifacts, Image Generation, Image Edit and Video Generation as product features. No paid consumer subscription and no numeric usage caps are published on the official pages fetched.

No numeric free allowance is published on official pages (checked 2026-09-16); qwen.ai states only that Qwen Studio is free to use and open to all.

Tencent AI Studio (aistudio.tencent.com, branded Hy AI Studio and built on Tencent Hunyuan) is Tencent's Chinese-language console for chatting with Hunyuan models, trying multimodal capabilities and managing API access. Tencent Cloud's official Hunyuan billing page states that after completing real-name verification, first activation issues a one-time free call allowance as a free resource pack: 1,000,000 tokens shared across Hunyuan-a13b, Hunyuan-role-latest, Hunyuan-translation, Hunyuan-translation-lite, Tencent HY Vision 1.5 Instruct, Hunyuan-turbos-vision, Hunyuan-t1-vision and Hunyuan-turbos-vision-video, plus a separate 1,000,000 tokens for Hunyuan-embedding, both valid one year. Postpaid list prices start at 0.5 yuan input and 2 yuan output per million tokens for Hunyuan-a13b.

Free grant is a one-time 1,000,000-token resource pack shared across the listed Hunyuan text and vision models, plus 1,000,000 tokens for Hunyuan-embedding, valid 1 year, stated in Tencent Cloud's Hunyuan billing documentation.

Tencent Yuanbao (yuanbao.tencent.com) — Tencent's free consumer AI assistant built on the Hunyuan (混元) family; current version serves Hunyuan Hy4 preview with upgraded Agent capabilities, plus a DeepSeek R1 inference option. Unofficial browser/cookie integration.

A free tier is available; the exact credit value is not quantified.