# WeightsAPI

> No KYC · No added content filters

Explore open-weight models, compare token prices and prepare an OpenAI-compatible inference workflow with WeightsAPI.

## Do I need to complete KYC?

The current sign-in flow has no identity-document upload or KYC verification step. A signed-in account is still required, and usage remains associated with that account.

- [Do I need to complete KYC?](https://weightsapi.com/faq#kyc-and-account)

## Are the models uncensored?

WeightsAPI adds no content filter at the gateway. The catalogue includes original models and fine-tuned variants published by their respective creators, including Dolphin and Hermes. Some variants are designed to reduce refusals; model training can still cause refusals. Read each model’s source and behavior notes.

- [Are the models uncensored?](https://weightsapi.com/faq#model-filters)

## Account access and model behavior

The current WeightsAPI account flow has no identity-document upload or KYC step; access and usage remain linked to your account. WeightsAPI adds no gateway content filter. The catalogue includes original models and published fine-tuned variants, including Dolphin and Hermes. Their training and configuration can still cause refusals.

- [Know what your request depends on.](https://weightsapi.com/trust#account-and-model-behavior)

## Account access

Sign in and confirm a top-up of at least USD 100, then spend your prepaid credit. Every recharge has the same minimum; smaller remaining balances stay usable. Check service status for model availability.

## Quickstart

- [llms-full.txt](https://weightsapi.com/llms-full.txt): Quickstart
- [service-info.json](https://weightsapi.com/service-info.json)
- [Managed LLM inference. Open models, one API.](https://weightsapi.com/): Explore open-weight models, compare token prices and prepare an OpenAI-compatible inference workflow with WeightsAPI.
- [Frequently asked questions](https://weightsapi.com/faq): Practical answers about WeightsAPI access, API integration, crypto credit, billing and data handling.
- [Know what your request depends on.](https://weightsapi.com/trust): Account controls, data handling, payment verification and the scope of service availability.
- [Model catalog](https://weightsapi.com/models): Choose the right balance of capability, context and cost.
- [Transparent by the token.](https://weightsapi.com/pricing): Compare pay-as-you-go token rates and prepaid monthly GPU rentals.
- [Quickstart](https://weightsapi.com/docs): Make your first OpenAI-compatible request.
- [Authentication](https://weightsapi.com/docs/authentication): Bearer keys, per-key budgets and model restrictions.
- [Billing and crypto deposits](https://weightsapi.com/docs/billing): Balance, confirmations, accounting and withdrawals.
- [Data handling](https://weightsapi.com/docs/privacy): What the gateway stores and what it never logs.
- [Service status](https://weightsapi.com/status): Model availability and incident history.
- [About WeightsAPI](https://weightsapi.com/about): How WeightsAPI approaches managed inference, shared and dedicated capacity, model choice and account controls.
- [What a million tokens actually costs](https://weightsapi.com/blog/cost-per-million): A million input tokens and a million output tokens are different products. With Qwen3 32B, 800,000 input and 200,000 output tokens cost $0.16 at the listed rates. The same total with an 80% output mix costs $0.19.
- [Estimate consumption before you migrate](https://weightsapi.com/blog/estimate-migration): Select a representative week and group requests by workflow. Preserve token counts and timings without keeping sensitive content in analytics.
- [Engineering notes](https://weightsapi.com/blog): Practical tradeoffs in running and using open models.
- [Latency and throughput answer different questions](https://weightsapi.com/blog/latency-throughput): TTFT includes queuing and prefill. It determines how long a user waits before seeing output.
- [Quantization is a tradeoff, not a free upgrade](https://weightsapi.com/blog/quantization): Lower precision reduces memory use and can improve throughput. Whether quality changes depends on the model, quantization method and task.
- [Self-hosting versus API: do the arithmetic](https://weightsapi.com/blog/self-hosting): Dedicated monthly cost divided by effective per-million-token API cost gives the volume crossover before engineering and reliability costs. An H100 at $1,690 and an effective $0.16 per million tokens gives about 10.56 billion tokens per month. This does not prove a single H100 can serve that workload.
- [Documentation and accounting console](https://weightsapi.com/changelog/2026-09-04-documentation-preview): Initial catalogue, documentation and account-management release, with deployment and payment availability recorded separately.
- [Changelog](https://weightsapi.com/changelog): Product changes and release notes, dated and traceable.
- [Compare models](https://weightsapi.com/compare): One source of truth for prices, context and capabilities.
- [Batch one hundred thousand documents](https://weightsapi.com/cookbook/batch-documents): Prepare a JSONL batch, submit it, then poll for the output file.
- [Stream to a browser](https://weightsapi.com/cookbook/browser-streaming): A same-origin Node server holds the API key and forwards SSE.
- [A chatbot in twenty lines](https://weightsapi.com/cookbook/chatbot): A terminal conversation using only the Python standard library.
- [Recipes you can run.](https://weightsapi.com/cookbook): Small, independent examples for real inference workflows.
- [Route by task difficulty](https://weightsapi.com/cookbook/model-routing): An explicit heuristic routes short tasks and complex analysis.
- [RAG with citations](https://weightsapi.com/cookbook/rag-with-citations): A tiny retrieval pipeline with verifiable source IDs.
- [Validated structured output](https://weightsapi.com/cookbook/structured-output): A strict schema plus independent validation.
- [An agent that calls a tool](https://weightsapi.com/cookbook/tool-agent): A bounded, allowlisted function call and a follow-up answer.
- [Rent a dedicated GPU for your model.](https://weightsapi.com/dedicated): Choose from six GPU configurations, place a monthly rental order and pay with your account balance or crypto credit.
- [API reference](https://weightsapi.com/docs/api-reference): Endpoints, request fields and response contracts.
- [Errors and recovery](https://weightsapi.com/docs/errors): Understand what failed and what to do next.
- [Function calling](https://weightsapi.com/docs/function-calling): Call allowlisted tools with validated arguments.
- [Rate limits and quotas](https://weightsapi.com/docs/limits): Token windows, funded account access and request limits.
- [Migrate from OpenAI or Anthropic](https://weightsapi.com/docs/migration): Exact changes for the Chat Completions API.
- [Streaming](https://weightsapi.com/docs/streaming): Consume server-sent events and handle disconnections.
- [Strict JSON output](https://weightsapi.com/docs/structured-output): Constrain shape and validate every response.
- [From a task to a clearer inference workflow.](https://weightsapi.com/how-it-works): Understand managed inference, choose a model, prepare a request and follow the relationship between credit, reserved amounts and usage.
- [Cline](https://weightsapi.com/integrations/cline): Configure Cline with the WeightsAPI OpenAI-compatible endpoint.
- [Continue.dev](https://weightsapi.com/integrations/continue): Configure Continue.dev with the WeightsAPI OpenAI-compatible endpoint.
- [Cursor](https://weightsapi.com/integrations/cursor): Configure Cursor with the WeightsAPI OpenAI-compatible endpoint.
- [Dify](https://weightsapi.com/integrations/dify): Configure Dify with the WeightsAPI OpenAI-compatible endpoint.
- [Your tools. Our inference.](https://weightsapi.com/integrations): Exact configuration guides for the tools you already use.
- [LangChain (Python)](https://weightsapi.com/integrations/langchain): Configure LangChain (Python) with the WeightsAPI OpenAI-compatible endpoint.
- [LibreChat](https://weightsapi.com/integrations/librechat): Configure LibreChat with the WeightsAPI OpenAI-compatible endpoint.
- [LlamaIndex (Python)](https://weightsapi.com/integrations/llamaindex): Configure LlamaIndex (Python) with the WeightsAPI OpenAI-compatible endpoint.
- [n8n](https://weightsapi.com/integrations/n8n): Configure n8n with the WeightsAPI OpenAI-compatible endpoint.
- [Obsidian](https://weightsapi.com/integrations/obsidian): Configure Obsidian with the WeightsAPI OpenAI-compatible endpoint.
- [Open WebUI](https://weightsapi.com/integrations/open-webui): Configure Open WebUI with the WeightsAPI OpenAI-compatible endpoint.
- [Roo Code](https://weightsapi.com/integrations/roo-code): Configure Roo Code with the WeightsAPI OpenAI-compatible endpoint.
- [SillyTavern](https://weightsapi.com/integrations/sillytavern): Configure SillyTavern with the WeightsAPI OpenAI-compatible endpoint.
- [DeepSeek R1](https://weightsapi.com/models/deepseek-r1): Difficult reasoning, mathematics and code analysis.
- [DeepSeek V3](https://weightsapi.com/models/deepseek-v3): Code generation, complex instructions and structured tasks.
- [DeepSeek V4 Flash](https://weightsapi.com/models/deepseek-v4-flash): A text model for long-context reasoning, coding and tool workflows, with a configured 1,048,576-token window.
- [Dolphin Mistral 24B Venice Edition](https://weightsapi.com/models/dolphin-mistral-24b-venice-edition): Conversations and characters with application-controlled instructions, from Dolphin and Venice.
- [Gemma 3 27B](https://weightsapi.com/models/gemma-3-27b): Image understanding and multilingual conversation.
- [Hermes 4 70B](https://weightsapi.com/models/hermes-4-70b): Creative conversation, roleplay and reasoning with configurable instructions, published by Nous Research.
- [Llama 3.1 405B](https://weightsapi.com/models/llama-3-1-405b): Demanding general-purpose generation and complex instructions.
- [Llama 3.1 8B](https://weightsapi.com/models/llama-3-1-8b): Fast classification, extraction and lightweight chat.
- [Llama 3.3 70B](https://weightsapi.com/models/llama-3-3-70b): General-purpose assistants, summarization and tool use.
- [Mistral Small 3](https://weightsapi.com/models/mistral-small-3): Efficient text tasks, function calling and European-language chat.
- [Mistral Small 4](https://weightsapi.com/models/mistral-small-4): An official Mistral model combining text and image input, tool calling and configurable reasoning, with a published 262,144-token window.
- [Mixtral 8×22B](https://weightsapi.com/models/mixtral-8x22b): Multilingual generation and function-calling workflows.
- [Qwen3 235B A22B](https://weightsapi.com/models/qwen3-235b-a22b): Strong general-purpose instruction following and tool use.
- [Qwen3 32B](https://weightsapi.com/models/qwen3-32b): A balanced choice for code, multilingual chat and reasoning.
- [Qwen3-Coder-Next](https://weightsapi.com/models/qwen3-coder-next): A text-only coding model for codebase exploration and tool workflows, with a published 262,144-token native context.
- [Privacy and data handling](https://weightsapi.com/privacy): Account records, request content and the scope of data handling.
- [Get a request back on track.](https://weightsapi.com/support): Diagnose API, streaming and billing problems, then prepare a useful report without sharing secrets.
- [Service conditions](https://weightsapi.com/terms): Payment, activation, rental periods and service conditions.
- [Use cases you can evaluate](https://weightsapi.com/use-cases): Practical scenarios for conversational assistants, code and tool workflows, and document questions, with clear inputs and quality checks.

## Language

- [English](https://weightsapi.com/llms.txt)
- [Français](https://weightsapi.com/fr/llms.txt)
- [Deutsch](https://weightsapi.com/de/llms.txt)
- [Español](https://weightsapi.com/es/llms.txt)
- [简体中文](https://weightsapi.com/zh/llms.txt)
- [Italiano](https://weightsapi.com/it/llms.txt)
- [Português (Brasil)](https://weightsapi.com/pt/llms.txt)
- [Nederlands](https://weightsapi.com/nl/llms.txt)
- [Polski](https://weightsapi.com/pl/llms.txt)
- [Русский](https://weightsapi.com/ru/llms.txt)
- [Українська](https://weightsapi.com/uk/llms.txt)
- [Türkçe](https://weightsapi.com/tr/llms.txt)
- [العربية](https://weightsapi.com/ar/llms.txt)
- [日本語](https://weightsapi.com/ja/llms.txt)
- [한국어](https://weightsapi.com/ko/llms.txt)