{"openapi":"3.0.0","info":{"description":"Hyperbolic serverless AI inference (OpenAI-compatible chat, text, image, and audio generation) plus GPU marketplace instance management; static Bearer key.","version":"v1"},"servers":[{"description":"Hyperbolic serverless AI inference (OpenAI-compatible chat, text, image, and audio generation) plus GPU marketplace instance management; static Bearer key.","url":"https://api.apirails.com/hyperbolic"}],"paths":{"/v1/models":{"get":{"tags":["Models"],"parameters":[],"responses":{"401":{"description":"Bearer token required (Authorization header)","content":{"application/json":{}}}},"externalDocs":{"api_ref":"https://docs.hyperbolic.ai/inference/overview","api_endpoint":"https://api.hyperbolic.xyz/v1/models"},"summary":"List available models","description":"Lists the models available on the OpenAI-compatible inference API. Endpoint live-verified (auth-gated) and anchored to the documented OpenAI SDK compatibility (client.models.list() works unchanged).","security":[{"bearerAuth":[]}]}},"/v1/chat/completions":{"post":{"tags":["Chat"],"parameters":[],"responses":{"401":{"description":"Bearer token required (Authorization header)","content":{"application/json":{}}}},"externalDocs":{"api_ref":"https://docs.hyperbolic.ai/inference/text-apis","api_endpoint":"https://api.hyperbolic.xyz/v1/chat/completions"},"summary":"Create chat completion","description":"The primary endpoint for text generation: chat completions with 20+ open-source models. Also used for vision-language models by passing image content parts.","security":[{"bearerAuth":[]}],"requestBody":{"description":"","required":true,"content":{"application/json":{"schema":{"type":"object","required":[],"properties":{"max_tokens":{"type":"number","description":"Maximum number of tokens to generate"},"messages":{"type":"array","description":"Array of message objects with role and content. Roles: system (sets behavior/context), user, assistant (conversation history). For vision models, content can be an array with image_url parts containing base64-encoded images.","required":true,"items":{"type":"object","properties":{"content":{"type":"string","description":"Content of the message","required":true},"role":{"type":"string","description":"Role of the message sender","required":true,"enum":["system","user","assistant"]}}}},"model":{"type":"string","description":"Model ID (e.g., deepseek-ai/DeepSeek-R1, meta-llama/Llama-3.3-70B-Instruct)","required":true},"stop":{"type":"array","description":"Stop sequence(s) to end generation. Send a single string or an array of strings.","items":{"type":"string"}},"stream":{"type":"boolean","description":"Enable streaming responses (default: false)","enum":[true,false]},"temperature":{"type":"number","description":"Sampling temperature (0-2). Higher = more creative. Docs example value 0.7."},"tool_choice":{"type":"string","description":"Controls tool use: \"auto\" lets the model decide, \"none\" disables tool use, or specify a tool name to force its use.","enum":["auto","none"]},"tools":{"type":"array","description":"Tool definitions for function calling. Supported by tool-calling models such as deepseek-ai/DeepSeek-R1, meta-llama/Llama-3.3-70B-Instruct, Qwen/Qwen3-235B-A22B-Instruct-2507, and openai/gpt-oss-120b.","items":{"type":"object","properties":{"type":{"type":"string","description":"Tool type","required":true,"enum":["function"]},"function":{"type":"object","description":"Function definition the model can call","required":true}}}},"top_p":{"type":"number","description":"Nucleus sampling threshold (0-1)"}}},"example":{"max_tokens":512,"messages":[{"content":"Explain quantum computing in simple terms.","role":"user"}],"model":"deepseek-ai/DeepSeek-R1","stop":["\n\nUser:"],"stream":"","temperature":0.7,"tool_choice":"auto","tools":[{"type":"function","function":{"description":"Get the current weather for a location","name":"get_weather","parameters":{"type":"object","properties":{"location":{"type":"string","description":"City and state, e.g. San Francisco, CA"},"unit":{"type":"string","description":"Temperature unit","enum":["celsius","fahrenheit"]}},"required":["location"]}}}],"top_p":1}}}}}},"/v1/completions":{"post":{"tags":["Completions"],"parameters":[],"responses":{"401":{"description":"Bearer token required (Authorization header)","content":{"application/json":{}}}},"externalDocs":{"api_ref":"https://docs.hyperbolic.ai/inference/text-apis","api_endpoint":"https://api.hyperbolic.xyz/v1/completions"},"summary":"Create text completion","description":"Raw text completion for base models (e.g., meta-llama/Meta-Llama-3.1-405B) without chat formatting; ideal for fill-in-the-middle tasks and custom prompting strategies.","security":[{"bearerAuth":[]}],"requestBody":{"description":"","required":true,"content":{"application/json":{"schema":{"type":"object","required":[],"properties":{"max_tokens":{"type":"number","description":"Maximum number of tokens to generate"},"model":{"type":"string","description":"Model ID of a base model (e.g., meta-llama/Meta-Llama-3.1-405B)","required":true},"prompt":{"type":"string","description":"The raw text prompt to complete","required":true},"temperature":{"type":"number","description":"Sampling temperature (0-2). Higher = more creative. Docs example value 0.7."}}},"example":{"max_tokens":256,"model":"meta-llama/Meta-Llama-3.1-405B","prompt":"The key principles of machine learning are","temperature":0.7}}}}}},"/v1/image/generation":{"post":{"tags":["Images"],"parameters":[],"responses":{"401":{"description":"Bearer token required (Authorization header)","content":{"application/json":{}}}},"externalDocs":{"api_ref":"https://docs.hyperbolic.ai/inference/image-apis","api_endpoint":"https://api.hyperbolic.xyz/v1/image/generation"},"summary":"Generate image","description":"Text-to-image generation with diffusion models (FLUX.1-dev, SDXL1.0-base, Stable Diffusion). The same endpoint performs image-to-image transformation (Stable Diffusion models only) when a base64 reference image is supplied.","security":[{"bearerAuth":[]}],"requestBody":{"description":"","required":true,"content":{"application/json":{"schema":{"type":"object","required":[],"properties":{"backend":{"type":"string","description":"Computation backend","enum":["auto","tvm","torch"]},"cfg_scale":{"type":"number","description":"Prompt relevance (higher = closer to prompt, typically 5-15)"},"enable_refiner":{"type":"boolean","description":"Enable the SDXL refiner for enhanced details","enum":[true,false]},"height":{"type":"number","description":"Image height in pixels"},"image":{"type":"string","description":"Base64-encoded reference image for image-to-image generation (Stable Diffusion models only; FLUX.1-dev does not support this feature)"},"model_name":{"type":"string","description":"Image model ID (e.g., FLUX.1-dev, SDXL1.0-base)","required":true},"negative_prompt":{"type":"string","description":"What to avoid in the generated image"},"negative_prompt_2":{"type":"string","description":"Secondary negative prompt for SDXL"},"prompt":{"type":"string","description":"Text description of the image to generate","required":true},"prompt_2":{"type":"string","description":"Secondary prompt for SDXL models"},"sampler":{"type":"string","description":"Sampling algorithm to use"},"seed":{"type":"number","description":"Random seed for reproducible results"},"steps":{"type":"number","description":"Number of inference steps (more = higher quality, slower)"},"strength":{"type":"number","description":"Image-to-image transformation strength (0-1). 0.3-0.5 subtle changes, 0.5-0.7 moderate, 0.7-1.0 major changes."},"width":{"type":"number","description":"Image width in pixels"}}},"example":{"backend":"auto","cfg_scale":5,"enable_refiner":"","height":1024,"image":"iVBORw0KGgoAAAANSUhEUgAA...","model_name":"FLUX.1-dev","negative_prompt":"blurry, low quality","negative_prompt_2":"oversaturated","prompt":"A panoramic view of the white sandy beach and turquoise sea","prompt_2":"cinematic lighting","sampler":"euler","seed":12345,"steps":30,"strength":0.7,"width":1024}}}}}},"/v1/audio/generation":{"post":{"tags":["Audio"],"parameters":[],"responses":{"401":{"description":"Bearer token required (Authorization header)","content":{"application/json":{}}}},"externalDocs":{"api_ref":"https://docs.hyperbolic.ai/inference/audio-apis","api_endpoint":"https://api.hyperbolic.xyz/v1/audio/generation"},"summary":"Generate speech from text","description":"Text-to-speech generation with Melo TTS. Returns a JSON object with base64-encoded MP3 audio in the \"audio\" field.","security":[{"bearerAuth":[]}],"requestBody":{"description":"","required":true,"content":{"application/json":{"schema":{"type":"object","required":[],"properties":{"text":{"type":"string","description":"The text to convert to speech","required":true},"language":{"type":"string","description":"Language code for speech generation (e.g., EN, EN-US, FR, JP, ZH)"},"noise_scale":{"type":"number","description":"Controls randomness in speech generation for more natural variation (0-1)"},"noise_scale_w":{"type":"number","description":"Controls randomness in duration prediction for timing variation (0-1)"},"sdp_ratio":{"type":"number","description":"Stochastic duration predictor ratio, controls prosody variation (0-1)"},"speaker":{"type":"string","description":"Speaker variant (language-specific)"},"speed":{"type":"number","description":"Speech speed multiplier (0.1-5)"}}},"example":{"text":"Welcome to Hyperbolic! This is an example of text-to-speech generation.","language":"EN","noise_scale":0.6,"noise_scale_w":0.8,"sdp_ratio":0.2,"speaker":"EN-Default","speed":1}}}}}},"/v1/marketplace/virtual-machine":{"post":{"tags":["GPU Marketplace"],"parameters":[],"responses":{"401":{"description":"Bearer token required (Authorization header)","content":{"application/json":{}}}},"externalDocs":{"api_ref":"https://docs.hyperbolic.ai/on-demand/quickstart","api_endpoint":"https://api.hyperbolic.xyz/v1/marketplace/virtual-machine"},"summary":"Create marketplace virtual machine","description":"Launch an on-demand GPU virtual-machine instance from the Hyperbolic marketplace. Billed per the console pricing; use deliberately.","security":[{"bearerAuth":[]}],"requestBody":{"description":"","required":true,"content":{"application/json":{"schema":{"type":"object","required":[],"properties":{"gpu_count":{"type":"number","description":"Number of GPUs for the instance","required":true},"gpu_type":{"type":"string","description":"GPU type to launch (e.g., H100, H200, B200)","required":true}}},"example":{"gpu_count":1,"gpu_type":"H100"}}}}}},"/v1/marketplace/bare-metal":{"post":{"tags":["GPU Marketplace"],"parameters":[],"responses":{"401":{"description":"Bearer token required (Authorization header)","content":{"application/json":{}}}},"externalDocs":{"api_ref":"https://docs.hyperbolic.ai/on-demand/quickstart","api_endpoint":"https://api.hyperbolic.xyz/v1/marketplace/bare-metal"},"summary":"Create marketplace bare metal node","description":"Launch a bare-metal GPU node from the Hyperbolic marketplace for multi-GPU workloads (e.g., 8x H100 with InfiniBand). Billed per the console pricing; use deliberately.","security":[{"bearerAuth":[]}],"requestBody":{"description":"","required":true,"content":{"application/json":{"schema":{"type":"object","required":[],"properties":{"gpu_count":{"type":"number","description":"Number of GPUs for the bare-metal node","required":true},"gpu_type":{"type":"string","description":"GPU type to launch (e.g., H100, H200)","required":true},"network_type":{"type":"string","description":"Interconnect network type for multi-GPU nodes"}}},"example":{"gpu_count":8,"gpu_type":"H100","network_type":"infiniband"}}}}}},"/v1/marketplace/instance/{id}":{"delete":{"tags":["GPU Marketplace"],"parameters":[{"name":"id","in":"path","description":"The marketplace instance ID to terminate","required":true,"schema":{"type":"string","example":""}}],"responses":{"401":{"description":"Bearer token required (Authorization header)","content":{"application/json":{}}}},"externalDocs":{"api_ref":"https://docs.hyperbolic.ai/on-demand/quickstart","api_endpoint":"https://api.hyperbolic.xyz/v1/marketplace/instance/{id}"},"summary":"Terminate marketplace instance","description":"Terminate (stop) a running marketplace instance by its instance ID. Destructive: unsaved work on the instance is lost.","security":[{"bearerAuth":[]}],"requestBody":{"description":"","required":true,"content":{"application/json":{"schema":{"type":"object","required":[],"properties":{}},"example":{}}}}}}},"components":{"securitySchemes":{"bearerAuth":{"type":"http","scheme":"bearer"}}},"security":[{"bearerAuth":[]}]}