Fresh third-party LLM API offers — auto-discovered twice daily, no credit card. Official big-name providers (Google, OpenAI, Anthropic, Groq, OpenRouter, Z.AI, Qwen, …) are excluded by design — their generic free quotas can't finish a real coding task. The bot hunts the new routers / aggregators / gateways that launch constantly, including short-lived 1–2 week free offers for frontier models.
- Aggressive search across 9+ sources:
- 🌐 Bing web search + SearXNG (multi-instance JSON)
- 📰 Google News RSS — daily free-offer news
- 🟠 Hacker News + 🔗 Lobste.rs
- 🐦 X/Twitter (Nitter instances) + Telegram public channels
- 🐙 GitHub repo search (newly launched providers)
- 🔴 Reddit (pullpush.io archive — datacenter-friendly)
- Lead scoring — every mention is keyword-scored (free / credits / trial / promo / giveaway / unlimited…).
- Auto-discovery ✨ — top leads are assessed (site live? speaks API? free/credit signal? not an official provider?):
- ✅ authentic → auto-added to
bot/providers.jsonand committed back to the repo
- ✅ authentic → auto-added to
- Daily link verification — every tracked provider's links get
OK / BLOCKED / DEAD / DOWNstatus. - Auto-clean — history capped at 90 entries, results fully regenerated each run, dead auto-added providers pruned.
- Open the table below (or the Pages site).
- Pick a fresh provider —
⚡= added this run,↻= auto-discovered earlier. - Click keys (or site), sign up with email, grab the API key.
- Point your OpenAI-format client at their base URL + key.
trust: low gateways = unofficial
resellers — never send sensitive data, expect flakiness, always check the actual free tier
terms before building on it.
| Path | Purpose |
|---|---|
bot/research.py |
The research bot (Python stdlib only, runs in GitHub Actions) |
bot/providers.json |
Third-party knowledge base + official-provider blocklist (bot auto-updates it) |
bot/queries.txt |
Daily search queries |
results/offers.json |
Latest structured results (machine-readable) |
results/leads.json |
Scored candidate leads from the latest run |
results/history.md |
Daily snapshot log (auto-capped) |
site/index.html |
GitHub Pages site (auto-generated) |
Actions → free-llm-daily-research → Run workflow (or gh workflow run research.yml).
- Runs only inside GitHub Actions — never touches the local machine.
results/+site/regenerated every run; history capped; dead auto-providers pruned; old Pages artifacts expire automatically.
Free tiers change fast; links are re-verified daily but nothing here is a guarantee. Auto-discovered providers are community finds, not endorsements.
6 providers verified OK | 0 flagged | 0 auto-discovered this run | last full run: 2026-09-08T10:34:42Z
| Provider | Client API | Models | Limits | Signup | Expiry | Links |
|---|---|---|---|---|---|---|
| LLM7.io (medium) | OPENAI-COMPAT | DeepSeek-R1; Qwen 2.5 | 30 RPM without signup; up to 5M tokens/day with free email token | no-CC none | ongoing | site free |
| BazaarLink (free gateway) (low) | OPENAI-COMPAT | Various via gateway (OpenAI-compatible) | Free OpenAI-compatible API key; model access varies | no-CC email | unknown | site free keys |
| yolo-auto.com ↻ (medium) | OPENAI-COMPAT | see website | Unmetered LLM API; paid plan reported ~$6/month, verify free tier | no-CC email | unknown | site free |
| openlegion.ai ↻ (medium) | OPENAI-COMPAT | see website | AI agent fleet with container isolation; free tier - verify | no-CC email | unknown | site free |
| komilion.com ↻ (medium) | OPENAI-COMPAT | see website | API router that picks the cheapest model for the task; free tier - verify | no-CC email | unknown | site free |
| blazephoenix.xyz ↻ (medium) | OPENAI-COMPAT | see website | see website | no-CC email | unknown | site free |
-
Free models you can use with your OpenClaw (no credit card needed) (news.ycombinator.com, via hackernews, score 7: free, api key, credit, trial, new, key, api)
-
Show HN: Faster more accurate multimodal vector search (github.com, via hackernews, score 6: free, api key, trial, week, key, api)
-
Show HN: Private AI assistant for $1.99 -Free AI (personalassistantdeploy.com, via hackernews, score 6: free, api key, api keys, month, key, api)
-
Google and OpenAI offer free API credits—but there's one catch - How-To Geek (news.google.com, via google-news, score 6: free, credits, credit, offer, new, api)
-
Show HN: One API for GPT-5, Claude-Sonnet-4, DeepSeek, Gemini (wisdom-gate.juheapi.com, via hackernews, score 3: aggregator, access, api)
-
Combining HTTP and JavaScript APIs with php (united-coders.com, via hackernews, score 3: free, aggregator, api)
-
Show HN: I made a free competition/giveaway aggregator (comps.gg, via hackernews, score 3: free, giveaway, aggregator)
-
Show HN: Recase – A free proxy server to cache API calls for you (recase.herokuapp.com, via hackernews, score 3: free, proxy, api)
-
Show HN: Spidra – AI web scraper that adapts to any website (spidra.io, via hackernews, score 3: new, launch, api)
-
Show HN: Pomo – Manage your Stripe promo codes without code (usecopi.com, via hackernews, score 3: free, promo, launch)
-
Show HN: LobsterLair – OpenClaw hosting with AI included ($19/mo) (lobsterlair.xyz, via hackernews, score 3: api key, key, api)
-
[TG:AI_Deals] Grab : Upto 85% Off on Aristocrat Trolleys + Bank Offers 😱 😱 🔗 Link: https://www.amazon.in/s?hidden-keywords (amazon.in, via telegram, score 3: offer, deal, key)
-
Free Online Games at Poki - Play Now! (bing.com, via bing, score 2: free, new)
-
Launch HN: Speko (YC S26) – OpenRouter for Voice AI (speko.ai, via hackernews, score 2: router, launch)
-
Show HN: LaunchPad – Job aggregator I built overnight after Amazon layoffs (launchpad-kappa-ashy.vercel.app, via hackernews, score 2: aggregator, launch)
-
Show HN: Smplogs – Local-first AWS Cloudwatch log analyzer via WASM (smplogs.com, via hackernews, score 2: gateway, api)
-
Show HN: AI coding agent for VS Code with pay-as-you-go pricing- no subscription (llmonestop.com, via hackernews, score 2: month, key)
-
Show HN: Get GPT-5.2, Grok-4.1-fast, KimiK2.5 and more LLMs at half the cost (frogapi.app, via hackernews, score 2: gateway, api)
-
Show HN: Till.sh, enhanced access controls for AWS S3 (till.sh, via hackernews, score 2: access, month)
-
Show HN: Zodii – a developer-first astrology, numerology and tarot API platform (zodiiapp.com, via hackernews, score 2: free, api)
-
Show HN: Twogether AI – Multi-Person Photo Generation API (twogether.ai, via hackernews, score 2: launch, api)
-
Show HN: Gensee – Free AI Agent Optimization and Deployment (platform.gensee.ai, via hackernews, score 2: free, launch)
-
Show HN: Quiltt Connector, Embeddable onboarding for finance products (quiltt.io, via hackernews, score 2: new, launch)
-
Show HN: Run any Llama model finetune and more, instantly (featherless.ai, via hackernews, score 2: new, api)
-
Show HN: SanctionSnap – free 250 sanctions checks via API (sanctionsnap.com, via hackernews, score 2: free, api)
-
Show HN: Gradient – a web API for fine-tuning and deploying Llama2 (gradient.ai, via hackernews, score 2: launch, api)
-
[TG:AI_Deals] MediBuddy Loot : Flat ₹5151 Off on Full Body Health Check-Up + FREE Vitamin Tests 🔥 🔥 👉 63+ Lab Tests (Vitam (bilty.co, via telegram, score 2: free, deal)
- bing: 12/12 queries returned results
- searxng: 0/8 queries returned results
- hackernews: 14/14 queries returned results
- reddit: 0/5 queries returned results
- github: 5/5 queries returned results
- google-news: 8/8 queries returned results
- x-twitter: 0/2 queries returned results
- telegram: 3 posts matched
- lobsters: 25 newest stories scanned
Last verified: 2026-09-08T10:34:42Z. Offers change fast - verify links before relying on them.