# SilentAPI > SilentAPI is a private API gateway for AI models and developer tools. One account and one API key give access to 61 models from Google AI, OpenAI, Anthropic, Meta AI, xAI, DeepSeek AI, Moonshot AI, Alibaba Cloud, Z.AI, Xiaomi, MiniMax, and Mistral through OpenAI, Anthropic, and Gemini compatible endpoints. SilentAPI does not keep prompts or outputs. Accounts are prepaid with crypto. Updated 2026-10-07. Key facts: - SilentAPI does not keep prompts, files, or outputs. Each model shows a privacy label: `End-to-end ZDR` (the model route has zero data retention) or `SilentAPI protected` (SilentAPI does not keep your prompts or outputs, and the model provider can keep logs under its own terms). - SilentAPI never uses a route whose provider explicitly states that it trains on customer data. SilentAPI does not log IP addresses with API requests. - Model providers do not see the IP address of a user, and they can not link a request to a SilentAPI account. - Account data is stored on servers in the European Union and handled in line with the GDPR. - Pricing is prepaid: you add credits with cryptocurrency through NOWPayments, or with a private payment, and each request is billed as you use it. One credit equals 0.01 USD, and there are no subscriptions and no hidden fees. - Crypto payments are anonymized. Each payment uses a random code that is deleted when the payment is complete, so a crypto wallet can not be tied to an account. Direct Monero deposits are also available. - Managed fallbacks: if a model or route fails, SilentAPI can retry on the next compatible model. - SilentAPI aims for 99.99% monthly uptime. The [status page](https://status.silentapi.com) shows live status and history. - Most open-weight language models run at 8-bit quantization, while many providers serve them at 4-bit quantization. Kimi models are an exception, and they can run at 4-bit quantization. The Services page and `silentapi.quantization` in the model list show the precision of each model. - SilentAPI systems are designed to the highest standards, with strict accuracy. The native APIs, such as lookups, media, and OCR, combine open source software, open data, licensed sources, and systems that SilentAPI builds, for extremely high accuracy. - Access over the public internet, Tor, and I2P. ## Start here - [Website](https://silentapi.com/): Product overview. - [Services and prices](https://silentapi.com/services): Live catalog with prices per model. - [Documentation](https://silentapi.com/docs/overview): API guides and examples. - [Model list](https://api.silentapi.com/public/models): Every model in OpenAI list format. No API key needed. - [API base URL](https://api.silentapi.com/v1): OpenAI and Anthropic style base URL. - [Status](https://status.silentapi.com): Live status with 24 hour, 30 day, and 1 year history. - [Tor API base URL](http://zwcbz3ftcbwhr2ercjghbozbqjj544mhwqd7guvjrxv2dkzwlq2prpyd.onion/v1): Tor access. - [I2P API base URL](http://ztkguqwo54vcgdgnji75k65exenh2hyznyfl4o7i5e3tin7padba.b32.i2p/v1): I2P access. ## Quick start ```bash curl https://api.silentapi.com/v1/chat/completions \ -H "Authorization: Bearer YOUR_SILENTAPI_KEY" \ -H "Content-Type: application/json" \ -d '{"model": "gemini-3.8-flash", "messages": [{"role": "user", "content": "Hello"}]}' ``` Use `Authorization: Bearer YOUR_SILENTAPI_KEY`. Do not put an API key in a URL, public file, client side application, or log. OpenAI SDKs work with `base_url` set to `https://api.silentapi.com/v1`. ## API styles - [OpenAI style](https://silentapi.com/docs/openai): Chat Completions API, Completions API, Responses API, embeddings, audio, images, model list, and model detail routes. - [Anthropic style](https://silentapi.com/docs/anthropic): `POST /v1/messages`. - [Gemini style](https://silentapi.com/docs/gemini): `POST /v1beta/models/{model}:generateContent`. - [Authentication](https://silentapi.com/docs/authentication): API keys, budgets, and expiry dates. - [Fallbacks](https://silentapi.com/docs/fallbacks): Retry on the next compatible model or route. - [Streaming](https://silentapi.com/docs/streaming): Server-sent events for all chat styles. - [Native services](https://silentapi.com/docs/overview): Lookup, media, OCR, moderation, music, finance, and status endpoints. ## Reasoning Chat models accept `reasoning_effort`. Each model supports its own levels, listed in `silentapi.reasoning` in the model list. SilentAPI changes an unsupported level to the closest supported level. ## Models The [Services page](https://silentapi.com/services) and `GET https://api.silentapi.com/public/models` are the current catalog. Prices are on the Services page and in `pricing` in the model list. ### Google AI - `gemini-3.1-pro`: Gemini 3.1 Pro. Chat, 1.05M context. End-to-end ZDR. - `gemini-3.8-flash`: Gemini 3.8 Flash. Chat, 1.05M context. End-to-end ZDR. - `gemini-3.5-flash-lite`: Gemini 3.5 Flash Lite. Chat, 1.05M context. End-to-end ZDR. - `gemini-3.8-flash-tts`: Gemini 3.8 Flash TTS. Text to speech, 8K context. SilentAPI protected. - `gemini-3.8-flash-lite-tts`: Gemini 3.8 Flash Lite TTS. Text to speech, 8K context. SilentAPI protected. - `gemini-3.5-transcribe`: Gemini 3.5 Transcribe. Speech to text, 98K context. SilentAPI protected. - `gemini-embedding-2`: Gemini Embedding 2. Embeddings, 8K context. End-to-end ZDR. - `lyria-3-pro`: Lyria 3 Pro. Music, 1.05M context. SilentAPI protected. - `lyria-3-clip`: Lyria 3 Clip. Music, 1.05M context. SilentAPI protected. - `gemini-3.1-flash-image`: Nano Banana 2. Image, 131K context. End-to-end ZDR. - `gemini-3.1-flash-lite-image`: Nano Banana 2 Lite. Image, 66K context. End-to-end ZDR. ### OpenAI - `gpt-6-astra`: GPT 6 Astra. Chat, 1.05M context. End-to-end ZDR. - `gpt-6.1-sol`: GPT 6.1 Sol. Chat, 1.05M context. End-to-end ZDR. - `gpt-6-sol`: GPT 6 Sol. Chat, 1.05M context. End-to-end ZDR. - `gpt-6-luna`: GPT 6 Luna. Chat, 1.05M context. End-to-end ZDR. - `gpt-5.6-sol`: GPT 5.6 Sol. Chat, 1.05M context. End-to-end ZDR. - `gpt-5.6-terra`: GPT 5.6 Terra. Chat, 1.05M context. End-to-end ZDR. - `gpt-5.6-luna`: GPT 5.6 Luna. Chat, 1.05M context. End-to-end ZDR. - `gpt-oss-120b`: GPT-OSS 120B. Chat, 131K context, 8-bit quantization. End-to-end ZDR. - `gpt-oss-20b`: GPT-OSS 20B. Chat, 131K context, 8-bit quantization. End-to-end ZDR. - `gpt-image-2.5-sunburst`: GPT Image 2.5 Sunburst. Image. SilentAPI protected. - `gpt-image-2.5-flare`: GPT Image 2.5 Flare. Image. SilentAPI protected. - `gpt-transcribe`: GPT Transcribe. Speech to text. SilentAPI protected. - `text-embedding-3-small`: Text Embedding 3 Small. Embeddings, 8K context. End-to-end ZDR. - `text-embedding-3-large`: Text Embedding 3 Large. Embeddings, 8K context. End-to-end ZDR. ### Anthropic - `claude-opus-5.5`: Claude Opus 5.5. Chat, 1M context. End-to-end ZDR. - `claude-opus-5`: Claude Opus 5. Chat, 1M context. End-to-end ZDR. - `claude-sonnet-5.5`: Claude Sonnet 5.5. Chat, 1M context. End-to-end ZDR. - `claude-sonnet-5`: Claude Sonnet 5. Chat, 1M context. End-to-end ZDR. - `claude-haiku-4.5`: Claude Haiku 4.5. Chat, 200K context. End-to-end ZDR. ### Meta AI - `muse-spark-1.3`: Muse Spark 1.3. Chat, 1.05M context. SilentAPI protected. - `muse-spark-1.2`: Muse Spark 1.2. Chat, 1.05M context. SilentAPI protected. ### xAI - `grok-4.7`: Grok 4.7. Chat, 500K context. End-to-end ZDR. - `grok-4.6`: Grok 4.6. Chat, 500K context. End-to-end ZDR. - `grok-imagine-image-2.0`: Grok Imagine Image 2.0. Image. SilentAPI protected. - `grok-imagine-video`: Grok Imagine Video. Video. SilentAPI protected. ### DeepSeek AI - `deepseek-v4.1-flash`: DeepSeek V4.1 Flash. Chat, 1.05M context, 8-bit quantization. End-to-end ZDR. - `deepseek-v4-flash`: DeepSeek V4 Flash. Chat, 1.05M context, 8-bit quantization. End-to-end ZDR. - `deepseek-v4-pro`: DeepSeek V4 Pro. Chat, 1.05M context, 8-bit quantization. End-to-end ZDR. ### Moonshot AI - `kimi-k3`: Kimi K3. Chat, 1.05M context, 4-bit quantization. End-to-end ZDR. - `kimi-k2.7-code`: Kimi K2.7 Code. Chat, 262K context, 4-bit quantization. End-to-end ZDR. - `kimi-k2.6`: Kimi K2.6. Chat, 262K context, 4-bit quantization. End-to-end ZDR. ### Alibaba Cloud - `qwen3.8-max`: Qwen3.8 Max (0902). Chat, 1M context. SilentAPI protected. - `qwen3.8-2.4t-a95b`: Qwen3.8 2.4T A95B. Chat, 1M context. End-to-end ZDR. - `qwen3.8-omni-flash`: Qwen3.8 Omni Flash. Chat, 1M context. SilentAPI protected. - `qwen3.8-flash`: Qwen3.8 Flash. Chat, 1M context. SilentAPI protected. - `qwen3-embedding-8b`: Qwen3 Embedding 8B. Embeddings, 33K context. End-to-end ZDR. ### Z.AI - `glm-5.3`: GLM 5.3. Chat, 1.05M context, 8-bit quantization. End-to-end ZDR. - `glm-5.2`: GLM 5.2. Chat, 1.05M context, 8-bit quantization. End-to-end ZDR. - `glm-5.3-flash`: GLM 5.3 Flash. Chat, 1.05M context, 8-bit quantization. End-to-end ZDR. ### Xiaomi - `mimo-v2.6-pro`: MiMo V2.6 Pro. Chat, 1.05M context, 8-bit quantization. End-to-end ZDR. - `mimo-v2.6-flash`: MiMo V2.6 Flash. Chat, 1.05M context, 8-bit quantization. End-to-end ZDR. ### MiniMax - `minimax-m3`: MiniMax M3. Chat, 1.05M context, 8-bit quantization. End-to-end ZDR. - `minimax-h3-max`: MiniMax H3 Max. Video. SilentAPI protected. - `minimax-h3`: MiniMax H3. Video. SilentAPI protected. ### Mistral - `mistral-medium-3.5`: Mistral Medium 3.5. Chat, 262K context. End-to-end ZDR. - `mistral-small`: Mistral Small 4. Chat, 262K context. End-to-end ZDR. - `mistral-embed-2312`: Mistral Embed 2312. Embeddings, 8K context. End-to-end ZDR. - `codestral-embed-2505`: Codestral Embed 2505. Embeddings, 8K context. End-to-end ZDR. - `voxtral-mini-transcribe`: Voxtral Mini Transcribe. Speech to text. End-to-end ZDR. - `voxtral-mini-tts`: Voxtral Mini TTS. Text to speech, 4K context. End-to-end ZDR. ## Media and native routes - [Media and OCR docs](https://silentapi.com/docs/media): Request formats and limits. - `POST /v1/images/compress`: Compress an image. - `POST /v1/images/convert`: Convert an image format. - `POST /v1/files/metadata/remove`: Remove metadata from images, PDF, and Office files. - `POST /v1/ocr`: OCR or Mistral OCR 4.1. - `POST /v1/moderations`: Mistral Moderation 2. - `POST /v1/audio/music`: Music generation. - `POST /v1/embeddings`: Embeddings. - `POST /v1/images/generations`: Image generation. - `POST /v1/videos`: Video jobs. - `POST /v1/audio/speech`: Text to speech. - `POST /v1/audio/transcriptions`: Speech to text. - `GET /v1/silentapi/ip-lookup`: IP lookup. - `GET /v1/silentapi/email-lookup`: Email lookup. - `GET /public/models`: Model list without an API key. - `GET /status`: Public status without an API key. Cached for 60 seconds. - `GET /status/history`: Public status history with 24 hourly, 30 daily, and 52 weekly buckets. ## Questions - Does SilentAPI store my prompts or outputs? No, SilentAPI does not keep prompts, files, or outputs. The privacy label of each model tells you what the model route keeps. - Which API formats work? OpenAI, Anthropic, and Gemini request formats all work, with streaming. - How does pricing work? It is a simple prepaid model: you add credits with cryptocurrency, and each request is billed as you use it. There are no subscriptions and no hidden fees. - How do I pay? You buy prepaid credits with cryptocurrency through NOWPayments, or with a private payment. - Which precision do open-weight models run at? Most run at 8-bit quantization, while many providers use 4-bit quantization. Kimi models are an exception, and they can run at 4-bit quantization. Each model line above shows its precision. - How accurate are the native APIs? They combine open source software, open data, licensed sources, and systems that SilentAPI builds and checks, for extremely high accuracy. - How do I find a model id? Call `GET https://api.silentapi.com/public/models`, or open the Services page. - Can I use SilentAPI over Tor or I2P? Yes, the site and the API are available on both networks. ## Comparisons - [All comparisons](https://silentapi.com/alternatives): Private AI services and AI API providers compared. - [SilentAPI alternatives](https://silentapi.com/alternatives/silentapi): The main alternatives to SilentAPI. - [OpenRouter alternatives](https://silentapi.com/alternatives/openrouter): SilentAPI and OpenRouter compared on signup, payment, Monero, logging, and Tor access. - [NanoGPT alternatives](https://silentapi.com/alternatives/nanogpt): SilentAPI and NanoGPT compared on signup, payment, Monero, logging, and Tor access. - [PayPerQ alternatives](https://silentapi.com/alternatives/payperq): SilentAPI and PayPerQ compared on signup, payment, Monero, logging, and Tor access. - [Venice.ai alternatives](https://silentapi.com/alternatives/venice): SilentAPI and Venice.ai compared on signup, payment, Monero, logging, and Tor access. - [Tinfoil alternatives](https://silentapi.com/alternatives/tinfoil): SilentAPI and Tinfoil compared on signup, payment, Monero, logging, and Tor access. - [Duck.ai alternatives](https://silentapi.com/alternatives/duck-ai): SilentAPI and Duck.ai compared on signup, payment, Monero, logging, and Tor access. - [Proton Lumo alternatives](https://silentapi.com/alternatives/proton-lumo): SilentAPI and Proton Lumo compared on signup, payment, Monero, logging, and Tor access. ## Optional - [Updates](https://silentapi.com/updates/0.9.9): Release notes. - [Payment methods](https://silentapi.com/payment): Prepaid account funding. - [Billing docs](https://silentapi.com/docs/billing): Credits, holds, and costs. - [Rate limits](https://silentapi.com/docs/rate-limits): Request limits. - [Errors](https://silentapi.com/docs/errors): Error codes. - [Privacy Policy](https://silentapi.com/privacy): Data handling. - [Terms](https://silentapi.com/terms): Service terms. - [Contact](https://silentapi.com/contact): Support and PGP key. - [Full documentation in one file](https://silentapi.com/llms-full.txt): All docs pages as Markdown.