What's new
By:

👨‍🏫 Tutorial Collection of Free AI API Providers

B

BELK

Honorary Poster
🤖 Collection of Free AI API Providers
Compiled list of free-tier AI routers/APIs — daily resets, free credits, and model access​

⚠️ Note: Providers requiring Telegram should be signed up with a separate Telegram account per provider to avoid cross-linking and spam.


🤖 Official / Direct Providers

1. Google AI Studio (Gemini)
You do not have permission to view the full content of this post. Log in or register now.
Free tier is model-specific; Google exposes RPM, input TPM, and RPD limits per model in AI Studio. Limits are not universal and can change.

  • Base URL: https://generativelanguage.googleapis.com/v1beta/openai/ — OpenAI-compatible; Anthropic-compatible: No.
  • Models: Gemini 2.5 Flash, Gemini 3 Flash, Gemini 3.5 Flash, Gemini 2.5 Flash Lite, Gemini 3.1 Flash Lite (~250k tokens/model)

2. Cloudflare Workers AI
You do not have permission to view the full content of this post. Log in or register now.
Free 10,000 Neurons/day (resets 00:00 UTC). Text generation is generally 300 requests/minute; model-specific limits apply. Some models require Workers ρáíd.

  • Base URL: https://api.cloudflare.com/client/v4/accounts/{account_id}/ai/v1 — OpenAI-compatible; Anthropic-compatible: No.
  • Models: OpenAI, Kimi, GLM, Qwen + more

3. Groq
You do not have permission to view the full content of this post. Log in or register now.
Free developer tier with model-specific RPM/RPD/TPM limits. Examples currently include Llama 3.1 8B: 30 RPM, 14,400 RPD, 6K TPM, 500K TPD; Llama 3.3 70B: 30 RPM, 1K RPD, 12K TPM, 100K TPD.

  • Base URL: https://api.groq.com/openai/v1 — OpenAI-compatible; Anthropic-compatible: No.
  • Models: Qwen, Llama, Gemma, Whisper + more

4. NVIDIA NIM
You do not have permission to view the full content of this post. Log in or register now.
Free NVIDIA NIM API access for development/prototyping. Limits vary by model, use case, and traffic; NVIDIA does not publish one universal quota. Current limits are shown in the API Catalog.

  • Base URL: https://integrate.api.nvidia.com/v1 — OpenAI-compatible; Anthropic-compatible: No.
  • Models: Mistral, DeepSeek, Qwen, Nemotron, Llama + more

5. Cohere
You do not have permission to view the full content of this post. Log in or register now.
Free trial API key: 20 Chat calls/minute, 5 Embed calls/minute, and 1,000 calls/month for other endpoints. Trial keys are for testing/prototyping and are not for production.

  • Base URL: https://api.cohere.ai/compatibility/v1 — OpenAI-compatible Compatibility API; Anthropic-compatible: No.
  • Models: Command, Command R, Command R+, Embed

6. Mistral AI
You do not have permission to view the full content of this post. Log in or register now.
Free mode is available by default for evaluation/prototyping. Limits are organization/model-specific: requests/sec, tokens/minute, and tokens/month; exact current limits are shown in the Mistral Admin Console.

  • Base URL: https://api.mistral.ai/v1 — OpenAI-compatible; Anthropic-compatible: No.
  • Models: Mistral Small, Mistral Medium, Ministral, Codestral

7. Logfare
You do not have permission to view the full content of this post. Log in or register now.
Free, no limits (data used for training unless opted out — opting out restricts to DeepSeek Flash only)

  • Models: Kimi, GLM, DeepSeek, Qwen, MiniMax, Kiro

8. Hcnsec
You do not have permission to view the full content of this post. Log in or register now.
Free ~4000 USD credits + daily check-in, unreliable, only use at night and use yolo/auto mode on coding agents, and use auto as model

  • Models: Kimi, GLM, DeepSeek, Qwen + more

9. OpenRouter
You do not have permission to view the full content of this post. Log in or register now.
Free 50 req/day (resets around 8AM), up to 1000 req/day with $10 credit top-up



🔀 Routers & Aggregators

10. Bluesmind
You do not have permission to view the full content of this post. Log in or register now.
Free $100

  • Requires GitHub account
  • Models: OpenAI, Anthropic (Claude), Gemini, Kimi, GLM, DeepSeek, Qwen + more (check model success rate)

11. Helpcoder
You do not have permission to view the full content of this post. Log in or register now.
Free $200 + daily check-in

  • Requires GitHub account
  • Models: OpenAI
  • Note: Failed when i tried testing

12. Agent Router
You do not have permission to view the full content of this post. Log in or register now.
Free $100 + $50 if refferal code is used + $25 daily check-in(log out, log in)

  • Requires GitHub account
  • Only works with coding agents (You do not have permission to view the full content of this post. Log in or register now.) — cannot be routed
  • Models: OpenAI, Claude, Use OpenAI models for now, Claude models have problems


🪙 Smaller / Misc Free Credits

13. Conduit
https://t(.)me()/conduitoff_bot?start=ref_8515810914
Free $15

  • Requires Telegram
  • Models: OpenAI, Claude, Grok, Kimi, Mistral, Meta

14. Nara Router
You do not have permission to view the full content of this post. Log in or register now.
Free 7M tokens/day

  • Requires Telegram
  • Models: Mistral, DeepSeek, Nemotron, Agnes, Laguna

15. QuickAITool
You do not have permission to view the full content of this post. Log in or register now.
Free credits

  • Requires Telegram
  • Models: OpenAI, Claude


16. Free Token Faucet
You do not have permission to view the full content of this post. Log in or register now.
Daily check-in free credits

  • Models: OpenAI, Claude, Kimi, GLM, DeepSeek, Qwen, MiniMax, Mimo

17. RiftAI
You do not have permission to view the full content of this post. Log in or register now.
Free 100k credits, resets daily

  • Models: OpenAI, Claude, Gemini, GLM, DeepSeek, Qwen

18. AI-Router
You do not have permission to view the full content of this post. Log in or register now.
Free $5 credit, $2/day usage limit, $1 daily check-in

  • Models: OpenAI, Claude

19. LivRouter
You do not have permission to view the full content of this post. Log in or register now.
Free $10 + $10 more via (D)(i)scord (tech support channel)

  • Models: OpenAI, Claude, Grok, GLM, DeepSeek

20. Kimchi Dev
You do not have permission to view the full content of this post. Log in or register now.

  • Use on 9router

21. Subaxis
You do not have permission to view the full content of this post. Log in or register now.
Free 1M tokens

  • Models: GLM 5.2

22. Aerolink
You do not have permission to view the full content of this post. Log in or register now.
Free 2-week trial

  • Limits: $4 / 5 hours, $27 / week
  • Models: GPT-5.6

23. Moon AI
You do not have permission to view the full content of this post. Log in or register now.
Free $180

  • They have weird pricing (~$4/minute regardless if you even use it or not)
  • Failed during testing, but you can try it yourself

24. Tabitoken
You do not have permission to view the full content of this post. Log in or register now.
Free $100 + $20 if invitation code used

  • Models: Claude Opus 4.8, Claude Opus 4.8 (Thinking), Claude Opus 5, Claude Opus 5 (Thinking)

25. ChatAnywhere
You do not have permission to view the full content of this post. Log in or register now.
Free 10,000 points/day, resets at 00:00

  • Models (5/day): GPT-5, GPT-5.1, GPT-5.2, GPT-5.4, GPT-5.5, GPT-5.6-Sol, GPT-5.6-Terra, GPT-4.1, GPT-4o
  • Models (100/day): GPT-5-mini, GPT-5-nano, GPT-5.4-mini, GPT-5.4-nano, GPT-5.6-Luna, GPT-4.1-mini, GPT-4.1-nano, GPT-4o-mini, GPT-3.5-turbo
  • Models (100/day): text-embedding-3-small, text-embedding-3-large, text-embedding-ada-002
  • Models (30/day): DeepSeek V3.2, V3.2-thinking, V4-Pro, V4-Flash, deepseek-chat
  • Once points run out, remaining requests are capped to 4096 max input tokens

26. Seekai
You do not have permission to view the full content of this post. Log in or register now.
Free $200

  • Models: Claude Fable 5, Claude Opus 4.7, Claude Opus 4.8, Claude Opus 5, Claude Sonnet 5, DeepSeek V4 Flash (0731), DeepSeek V4 Flash, DeepSeek V4 Pro, DeepSeek V4 Pro (0813), Gemini 3.1 Pro, Gemini 3 Flash, Gemini 3 Pro, GLM 5.2, GPT-5.4, GPT-5.5, GPT-5.6, GPT-5.6-Sol, GPT-5.6-Luna, GPT-5.6-Terra, Grok 4.5

27. GoRouter
You do not have permission to view the full content of this post. Log in or register now.
Free $50 + $20 if invitation code is used

  • Models: Claude Opus 4.8, Claude Opus 4.8 (Thinking), Claude Opus 5, Claude Opus 5 (Thinking)


☁️ Ollama Cloud (Official)

Ollama Cloud
You do not have permission to view the full content of this post. Log in or register now.

The following models currently work without an Ollama Cloud subscription (audit: 2026-07-20):

  • GPT-OSS 120B
  • GPT-OSS 20B
  • Nemotron 3 Ultra
  • Nemotron 3 Super
  • MiniMax M3
  • MiniMax M2.5
  • Gemma 4 31B
  • Nemotron 3 Nano
  • Qwen 2.5 1.5B

Subscription currently required for DeepSeek V4 Flash/Pro, GLM 5.1/5.2, Kimi K2.x, MiniMax M2.7, Mistral Large 3, and Qwen 3.5 397B.


🔀 Recommended: Use a Router/Aggregator

Instead of juggling separate API keys and endpoints for each provider above, point them all through a self-hosted or third-party router. One OpenAI-compatible endpoint, automatic fallback when a provider's free quota runs out, and unified usage tracking.

Bifrost
You do not have permission to view the full content of this post. Log in or register now.
B

👨‍🏫 Tutorial Beginner Guide to Bifrost


​

Bifrost AI Gateway: Beginner Installation and Usage Guide​


A practical introduction to installing, configuring, and using You do not have permission to view the full content of this post. Log in or register now. as a local AI gateway.

What is Bifrost?​


Bifrost is an AI gateway that provides one OpenAI-compatible API for multiple model providers. Instead of integrating every application directly with OpenAI, Anthropic, Google, AWS Bedrock, or another provider, an application can send requests to Bifrost.

Bifrost also provides a web interface for provider configuration, request monitoring, and analytics.

What you will need​


  • ...
  • Love
Self-hosted, open-source LLM gateway. OpenAI-compatible API, supports Gemini/OpenRouter/Cloudflare/etc. as providers, built-in Prometheus metrics, key management via REST API. Best option if you already run a homelab/VPS.

9router
You do not have permission to view the full content of this post. Log in or register now.
Hosted router — plug in Cloudflare Workers AI, Kimchi Dev, and similar free-tier providers behind one key.

https://phcorner.org/threads/2558506

Omnirouter
Similar concept to 9router — aggregates multiple free-tier providers (e.g. Cloudflare Workers AI) behind a single unified endpoint.

Why bother:

  • One API key/endpoint in your app instead of N
  • Automatic failover when a provider hits its daily free-tier cap
  • Centralized logging/metrics (token usage, success rate per provider)
  • Easier to swap/rotate providers without touching app code


🛠️ Using These APIs With Coding Agents

Most providers in this list expose an OpenAI-compatible API. The exact setup depends on whether your client expects OpenAI or Anthropic's API format.

⚠️ Important: A provider being marked OpenAI-compatible does not automatically mean it works with Claude Code or Codex. Claude Code needs an Anthropic-compatible endpoint, while current Codex custom providers use the OpenAI Responses API. OpenAI Chat Completions-only endpoints may work with OpenCode/Cline but not necessarily with Codex.

1. Claude Code

Claude Code uses Anthropic's API format. If your provider/router exposes an Anthropic-compatible endpoint, set the Anthropic base URL and API key before starting Claude Code.

  • Linux/macOS: set ANTHROPIC_BASE_URL and ANTHROPIC_API_KEY in your shell.
  • Windows PowerShell: use $env:ANTHROPIC_BASE_URL and $env:ANTHROPIC_API_KEY.
  • Use the provider's Anthropic-compatible base URL, not its OpenAI-compatible URL.

Bash: 
export ANTHROPIC_BASE_URL="https://YOUR-PROVIDER/anthropic"
export ANTHROPIC_API_KEY="YOUR_API_KEY"
claude

Code: 
$env:ANTHROPIC_BASE_URL="https://YOUR-PROVIDER/anthropic"
$env:ANTHROPIC_API_KEY="YOUR_API_KEY"
claude

If a provider only has an OpenAI-compatible endpoint, do not put that URL in ANTHROPIC_BASE_URL unless the provider specifically documents Anthropic compatibility.

2. Codex CLI

Codex supports custom model providers in ~/.codex/config.toml. Current Codex provider configuration uses an OpenAI-compatible Responses API endpoint. Set wire_api = "responses".

Code: 
model = "YOUR_MODEL_ID"
model_provider = "myprovider"

[model_providers.myprovider]
name = "My Provider"
base_url = "https://YOUR-PROVIDER/v1"
env_key = "MY_PROVIDER_API_KEY"
wire_api = "responses"

Then set the API key:

Bash: 
export MY_PROVIDER_API_KEY="YOUR_API_KEY"
codex

Code: 
$env:MY_PROVIDER_API_KEY="YOUR_API_KEY"
codex

Codex note: The provider must support the Responses API. A provider that only implements /v1/chat/completions is not suitable for current Codex custom-provider configuration.

3. OpenCode

OpenCode can use any OpenAI-compatible provider. Add the provider to opencode.json and use the provider's OpenAI-compatible base URL.

JSON: 
{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "myprovider": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "My Provider",
      "options": {
        "baseURL": "https://YOUR-PROVIDER/v1",
        "apiKey": "{env:MY_PROVIDER_API_KEY}"
      },
      "models": {
        "YOUR_MODEL_ID": {
          "name": "Your Model"
        }
      }
    }
  }
}

Set the key:

Bash: 
export MY_PROVIDER_API_KEY="YOUR_API_KEY"
opencode

Code: 
$env:MY_PROVIDER_API_KEY="YOUR_API_KEY"
opencode

  • For standard OpenAI-compatible /v1/chat/completions providers, use @ai-sdk/openai-compatible.
  • Replace YOUR_MODEL_ID with the exact model ID exposed by the provider.
  • OpenCode also supports custom baseURL values for proxies and gateways.

4. Cline in VS Code

Cline supports OpenAI-compatible providers directly from its settings.

  1. Open VS Code and open the Cline panel.
  2. Click the ⚙️ Settings button.
  3. Set API Provider to OpenAI Compatible.
  4. Enter the provider's Base URL, for example https://YOUR-PROVIDER/v1.
  5. Enter your API Key.
  6. Enter the exact Model ID supplied by the provider.
  7. Save/verify the configuration and send a simple test prompt.

Do not automatically append another /v1 if the provider's Base URL already includes /v1. Use the exact base URL shown in the provider entry above.

🔄 Quick Compatibility Guide

  • Claude Code → Anthropic-compatible endpoint
  • Codex → OpenAI Responses-compatible endpoint
  • OpenCode → OpenAI-compatible endpoint
  • Cline → OpenAI-compatible endpoint

💡 Using a Router/Gateway

If you use a gateway such as Bifrost, LiteLLM, or another router, the gateway can provide the compatibility layer required by the client. Configure the client against the gateway's appropriate endpoint instead of the upstream provider's endpoint.

  • Claude Code: gateway's Anthropic-compatible endpoint
  • Codex: gateway's OpenAI Responses-compatible endpoint
  • OpenCode/Cline: gateway's OpenAI-compatible endpoint
 

Similar threads

About this Thread

  • 6
    Replies
  • 897
    Views
  • 7
    Participants
Last reply from:
deadskull

New Topics

Trending Topics

Online now

Members online
1,223
Guests online
2,387
Total visitors
3,610

Forum statistics

Threads
2,318,002
Posts
29,193,838
Members
1,180,757
Latest member
cal2
Back
Top