SoftHive.ru

О проекте SoftHive.ru

SoftHive.ru — открытый каталог нейросетей и агрегатор бесплатных LLM API. Мы собираем, тестируем и сравниваем модели от Google AI Studio, Groq, DeepSeek, OpenRouter, Mistral, Cohere, Cerebras, Cloudflare Workers AI и десятков других провайдеров, давая разработчикам прозрачные лимиты запросов, фильтры работы без VPN в РФ и готовые конфиги для Cursor и Claude Code.

Why We Built This

There are hundreds of free LLM APIs available across the globe, but no single place to compare them side by side. Developers waste hours reading docs, registering accounts, and testing endpoints just to find a capable free coding model. SoftHive.ru puts all the information — context windows, rate limits, modality, API format compatibility — in one searchable, filterable interface with instant config generators.

Contribute

This project is backed by the open community on GitHub. Found a missing provider, broken endpoint, or incorrect parameter? We welcome contributions, PRs, and reports.

View GitHub Repository↗Follow on X / Twitter↗

How It Works

Our platform tracks free LLM models across 30+ providers. We monitor which models are online, which have gone offline, and whether rate limits or free-tier quotas have changed. The catalog rebuilds with live benchmark signals (MMLU-Pro, HumanEval, Math, Arena Elo) so you always know model quality before sending tokens.

Editorial Team & Review Process

Articles and directory guides are published and reviewed by the FreeLLM Hub Editorial Team. We cross-check provider claims against official documentation, live endpoint tests, and our public model catalog before publication.

Time-sensitive facts such as availability, free-tier status, and quotas are reviewed continuously. Each guide clearly indicates its verified status so readers can evaluate freshness.

How We Verify Models — Manually Tested

Most free LLM directories scrape lists without testing whether the APIs actually respond. We do things differently — models are validated via live API calls before being confirmed as online.

Live API testingWe send real chat completion requests using valid API keys to verify endpoint status, latency, and streaming support.
Error diagnosisWhen an endpoint fails (404, 403, card required, regional blocks), we identify root causes and adjust model IDs (e.g. correcting dash vs slash formats).
Status accuracyModels that consistently fail or exhaust free quotas are marked as offline or rate-limited.
Automated refreshesScripts regularly synchronize provider status and surface new models.

Our Principles

  • 01No fictional data. Every model on SoftHive.ru is verified against real provider responses. We don't fabricate models or fake benchmarks.
  • 02No credit card gate. We highlight which providers require zero payment card credentials so you can start developing in seconds.
  • 03Direct links, no reselling. We link you directly to the official provider signup. We never resell tokens or act as an intermediary man-in-the-middle proxy.
  • 04Data freshness. Model statuses evolve rapidly; we clearly display verification indicators and timestamps.

Contact & Support

If you are a model provider looking to update your listings, or have business inquiries, sponsorship questions, or general feedback, contact us directly:

Telegram: @softhive_ru