FreeLLM directory
A community free LLM API directory: awesome-freellm-apis (data refreshes daily from freellm.net). Dozens of providers, hundreds of free models, with notes on credit cards, phone verification, Base URL, and Key links.
This page is a MiniTavern user summary: how to pick a vendor from the directory, and how it maps to dedicated pages and the overview. Model IDs, rate limits, and whether something is still free follow freellm.net / each vendor console (the repo README notes the last update date).
Sites and entry points
| Resource | URL |
|---|---|
| GitHub repo | open-free-llm-api/awesome-freellm-apis |
| Chinese README | README.zh-CN.md |
| Online site | freellm.net (browse · Playground · config generator) |
| API Key directory | freellm.net/free-llm-api-keys |
| Model compare | freellm.net/compare |
How to use (with MiniTavern)
- On freellm.net or the repo “provider directory”, pick a vendor: prefer no credit card (e.g. Groq, Gemini, Zhipu Flash).
- Tap the provider’s Get Key, register, and copy the API Key.
- Note Base URL and Model ID (repo Quick Reference / Best Models tables).
- Open MiniTavern: Settings → Models → Model link → New
- If this handbook already has a dedicated page / App preset (table below) → follow that page.
- Otherwise pick Custom (OpenAI), paste Base URL + Key + model ID → Test.
General steps: Model link. For roleplay, prefer chat-oriented text models; coding Agent models may not suit long stories.
| Item | Details |
|---|---|
| Inclusion criteria (official) | Clear free tier (not only short trials), publicly reachable; trial credits must be labeled |
| Freshness | Auto-updated daily; limits and lists change |
| vs this handbook | Overview and dedicated pages cover “how to fill the App”; this page + freellm.net cover “what else you can try” |
Dedicated pages in this handbook (read first)
| Provider | Handbook | App platform |
|---|---|---|
| Agnes AI | Agnes AI | Custom (OpenAI) / Images |
| OpenCode Zen | OpenCode Zen | Custom (OpenAI) |
| AIHubMix | AIHubMix | Custom (OpenAI) / Images |
| OpenRouter | OpenRouter | OpenRouter |
| Zhipu / SiliconFlow / Bailian / DeepSeek / Gemini / Groq, etc. | Free cloud overview | See each overview section |
Provider directory (permanent free tiers · summary)
Excerpt from the repo “Permanent Free Tiers” (~2026-08-20). Free model counts change; full table: GitHub / freellm.net.
| Provider | ~Free models | Card etc. | Capabilities | Key |
|---|---|---|---|---|
| NVIDIA NIM | 125 | Phone verification | Text / image / video / vision, etc. | build.nvidia.com |
| ModelScope | 56 | Sign-up | Text / image / video / vision | modelscope.cn |
| Cloudflare Workers AI | 40 | No | Text / image / code, etc. | dash.cloudflare.com |
| Google Gemini | 17 | No | Text / multimodal | aistudio.google.com |
| GitHub Models | 16 | No | Text, etc. | GitHub Models |
| LLM7.io | 16 | No | Text / multimodal | token.llm7.io |
| OVHcloud AI Endpoints | 14 | Sign-up | Text, etc. | ovhcloud.com |
| Ollama Cloud | 13 | Sign-up | Text / vision, etc. | ollama.com/settings/keys |
| Groq | 12 | No | Text | console.groq.com/keys |
| Mistral AI | 12 | No | Text | console.mistral.ai |
| Cohere | 12 | No | Text | dashboard.cohere.com |
| Kilo Code | 12 | No | Text, etc. | kilo.ai |
| OpenCode Zen | 11 | Sign-up | See dedicated page | opencode.ai/auth |
| Cerebras | 8 | No | Text | cloud.cerebras.ai |
| Aion Labs | 7 | Sign-up | Text (incl. RP-oriented) | aionlabs.ai |
| Hugging Face | 7 | No | Text / code | huggingface.co/settings/tokens |
| Agnes AI | 5 | Sign-up | See dedicated page | platform.agnes-ai.com |
| Alibaba Cloud Bailian (intl, etc.) | 5 | Sign-up | See overview Qwen | Bailian console |
| Zhipu Z AI | 4 | No | See overview Zhipu | open.bigmodel.cn |
| SambaNova | 4 | Sign-up | Text | cloud.sambanova.ai |
| SiliconFlow | 3 | Sign-up | See overview SiliconFlow | cloud.siliconflow.cn |
| xAI / Grok | 3 | Sign-up | Text (often credits) | console.x.ai |
| DeepSeek | 2 | Sign-up | See overview DeepSeek | platform.deepseek.com |
| Others (Chutes, Glhf, AI21, Nscale, Nebius, etc.) | Few | Mostly sign-up | Follow freellm | See repo Quick Reference |
Renewable free quotas
| Provider | Notes | Handbook |
|---|---|---|
| OpenRouter | :free pool; ~$10+ top-up can raise free daily request caps | OpenRouter |
Quick reference · Base URL and App
Most vendors work with Custom (OpenAI). For Gemini use the App preset Gemini(Google) — do not force an OpenAI Base URL.
| Provider | Base URL (common) | App |
|---|---|---|
| Groq | https://api.groq.com/openai/v1 | Custom (OpenAI) |
| SiliconFlow | https://api.siliconflow.cn/v1 | Custom (OpenAI) |
| Zhipu | https://open.bigmodel.cn/api/paas/v4 | Custom (OpenAI) |
| Agnes AI | https://apihub.agnes-ai.com/v1 | Custom (OpenAI) |
| OpenCode Zen | https://opencode.ai/zen/v1 | Custom (OpenAI) |
| OpenRouter | https://openrouter.ai/api/v1 | OpenRouter |
| NVIDIA NIM | https://integrate.api.nvidia.com/v1 | Custom (OpenAI) |
| ModelScope | https://api-inference.modelscope.cn/v1 | Custom (OpenAI) |
| GitHub Models | https://models.github.ai/inference | Custom (OpenAI) |
| Cerebras | https://api.cerebras.ai/v1 | Custom (OpenAI) |
| SambaNova | https://api.sambanova.ai/v1 | Custom (OpenAI) |
| Mistral | https://api.mistral.ai/v1 | Custom (OpenAI) |
| Hugging Face | https://router.huggingface.co/v1 | Custom (OpenAI) |
| LLM7.io | https://api.llm7.io/v1 | Custom (OpenAI) |
| Ollama Cloud | https://api.ollama.com (follow docs) | Custom (OpenAI) |
| Gemini | Preset | Gemini(Google) |
| DeepSeek | Preset | DeepSeek |
| Qwen / Bailian | Preset | Qwen |
Cloudflare Workers AI paths include {account_id} and are fiddly; if unsure, start with Groq / Zhipu / OpenRouter.
Full Key links and “best free model per vendor” tables: repo Best Free Models.
Popular free models (repo weekly usage examples)
Only “recently used” references — not necessarily best for roleplay:
| Direction | Examples (follow freellm / plaza IDs) |
|---|---|
| Aggregated free pool | OpenRouter: nvidia/nemotron-3-ultra-550b-a55b:free, etc. |
| Fast inference | Groq: llama-3.3-70b-versatile, moonshotai/kimi-k2-instruct |
| Chinese chat | Zhipu Flash, SiliconFlow free-tagged models, Bailian new-user quota |
| Multimodal text + image gen | Agnes AI, AIHubMix |
Images / video
- Many directory rows mark
image/videomodality; MiniTavern image gen needs OpenAI Images–compatible or App-supported platforms (see Free image gen). - The App has no built-in video generation; video models are only for comparing site capabilities.