# HIPS

> HIPS is building a centrally managed AI inference marketplace on a distributed GPU network. It connects people buying inference with independently supplied GPU time.

We are inviting early-access interest while we develop the platform. Joining the list does not create an operating inference account or guarantee an invitation date.

## Pages
- Home: https://hips.com/ : HIPS is building managed inference for open-weight and customer-owned models. Join early access for updates on availability.
- AI inference API for your applications: https://hips.com/inference : Explore the HIPS approach to managed model inference, usage-based pricing, and always-on capacity. Register early-access interest.
- Open-weight AI models for inference: https://hips.com/inference/models : Discover the planned HIPS model catalog across language, image, audio, video, and specialist tasks.
- Deploy your AI model for inference: https://hips.com/inference/deploy : Learn how customer-owned models fit into HIPS, including execution-pool choice, data policies, and always-on capacity.
- Always-on GPU capacity for AI inference: https://hips.com/inference/always-on : Explore hourly capacity designed to keep supported models loaded between requests. Join HIPS early access.
- More compute put to work: https://hips.com/inference/how-it-works
- Understand how your models and data will be used: https://hips.com/inference/security
- AI inference for assistants and agents: https://hips.com/inference/use-cases/assistants-and-agents
- AI inference for document processing: https://hips.com/inference/use-cases/document-processing
- AI inference for image and video: https://hips.com/inference/use-cases/image-and-video
- AI inference for speech and audio: https://hips.com/inference/use-cases/speech-and-audio
- AI inference for classification and content analysis: https://hips.com/inference/use-cases/classification-and-content-analysis
- AI inference for software and SaaS: https://hips.com/inference/use-cases/software-and-saas : Build assistants, document features, and specialist model workflows with HIPS. Join early access for managed inference.
- AI inference for ecommerce and retail: https://hips.com/inference/use-cases/ecommerce-and-retail : Explore AI workflows for catalog information, product imagery, and customer support. Register HIPS early-access interest.
- AI inference for creative and media workflows: https://hips.com/inference/use-cases/creative-and-media : Explore image, video, audio, and language models for creative applications. Join HIPS early access.
- Earn with your GPU: https://hips.com/earn : Register interest in contributing available GPU capacity to HIPS, with schedules, price controls, and earnings visibility.
- AI inference pricing that follows the work: https://hips.com/pricing
- More room to build: https://hips.com/company/about
- Resources: https://hips.com/company/resources : Docs, Security and Data, Legal, and the machine-readable files
- Contact HIPS: https://hips.com/company/contact : Talk about inference, GPU capacity, or a project, and find the separate HIPS Payments path for existing customers.
- Affiliate program: https://hips.com/affiliates : How the HIPS affiliate program works: your referral link and code, the discount for people you refer, commissions, holds, payouts and the program rules. (The affiliate program opens later.)

## Docs
- An AI inference API for developers: https://hips.com/docs : A familiar integration path for open-weight language models and the HIPS SDK for other model types. These pages preview the intended experience. (preview)
- Your first HIPS inference request: https://hips.com/docs/quickstart : The intended setup for a supported chat-completion model using familiar clients, in Python, JavaScript, or cURL. (preview)
- Move your inference API to HIPS: https://hips.com/docs/migrate : Check the features your application uses, test quality and operating behavior, and move deliberately. (preview)
- The HIPS SDK for other supported model types: https://hips.com/docs/sdk : The intended path for image, video, audio, and specialist models such as classifiers. (preview)
- Import and run your model: https://hips.com/docs/import-and-run : From a model on Hugging Face to inference: import, compatibility checks, profiling, pool choice, and policies. (preview)
- Models and task guides: https://hips.com/docs/models : What each model accepts, returns, and costs, and each task's request format, usage units, and result handling. (preview)
- Task guides: https://hips.com/docs/tasks : Per task, the request format, the usage unit and the result handling, in that order. (preview)
- Language (chat): https://hips.com/docs/tasks/chat : Language (chat) on HIPS: the request format, the usage unit and the result handling. (preview)
- Classification: https://hips.com/docs/tasks/classification : Classification on HIPS: the request format, the usage unit and the result handling. (preview)
- Image: https://hips.com/docs/tasks/image : Image on HIPS: the request format, the usage unit and the result handling. (preview)
- Video: https://hips.com/docs/tasks/video : Video on HIPS: the request format, the usage unit and the result handling. (preview)
- Audio: https://hips.com/docs/tasks/audio : Audio on HIPS: the request format, the usage unit and the result handling. (preview)
- Document processing: https://hips.com/docs/tasks/document-processing : Document processing on HIPS: the request format, the usage unit and the result handling. (preview)
- Coding-tool setup: https://hips.com/docs/tools : Supported configurations for coding tools, each validated separately before it is listed as supported. (preview)
- Claude Code: https://hips.com/docs/tools/claude-code : Claude Code with HIPS: status, versions tested, what changes, how credentials are handled, how to restore the previous setup. (preview)
- Other coding tools and editor integrations: https://hips.com/docs/tools/openai-style-tools : Other coding tools and editor integrations with HIPS: status, versions tested, what changes, how credentials are handled, how to restore the previous setup. (preview)
- The HIPS SDK for Python: https://hips.com/docs/sdk/languages/python : Install and use the HIPS SDK for Python: configuration, a first task, jobs, errors and retries. (preview)
- The HIPS SDK for TypeScript and JavaScript: https://hips.com/docs/sdk/languages/node : Install and use the HIPS SDK for TypeScript and JavaScript: configuration, a first task, jobs, errors and retries. (preview)
- The HIPS SDK for Go: https://hips.com/docs/sdk/languages/go : Install and use the HIPS SDK for Go: configuration, a first task, jobs, errors and retries. (preview)
- The HIPS SDK for Rust: https://hips.com/docs/sdk/languages/rust : Install and use the HIPS SDK for Rust: configuration, a first task, jobs, errors and retries. (preview)
- The HIPS SDK for Java (usable from Kotlin): https://hips.com/docs/sdk/languages/java : Install and use the HIPS SDK for Java (usable from Kotlin): configuration, a first task, jobs, errors and retries. (preview)
- The HIPS SDK for .NET: https://hips.com/docs/sdk/languages/dotnet : Install and use the HIPS SDK for .NET: configuration, a first task, jobs, errors and retries. (preview)
- The HIPS SDK for PHP: https://hips.com/docs/sdk/languages/php : Install and use the HIPS SDK for PHP: configuration, a first task, jobs, errors and retries. (preview)
- The HIPS SDK for Ruby: https://hips.com/docs/sdk/languages/ruby : Install and use the HIPS SDK for Ruby: configuration, a first task, jobs, errors and retries. (preview)
- Batch jobs: https://hips.com/docs/batch : Send many requests at once, run them when capacity is free at half the on-demand price, and collect the results when they are done. (preview)
- Prompt caching: https://hips.com/docs/prompt-caching : Automatic prompt caching: cached input tokens at their own lower rate, no write charge, isolated per workspace. (preview)
- The hi command line: https://hips.com/docs/cli : hi, the HIPS command line for buyers: sign-in, models, batches, imports, keys and usage, with no API key. (preview)
- Sign in with device authorization: https://hips.com/docs/cli/login : How hi and HIPS Code sign in: a code, a link and a QR code in the terminal, approved in the Console, with no API key. (preview)
- HIPS API reference: https://hips.com/docs/reference : The HIPS inference API routes, parameters, headers, errors, and usage fields, generated from the contract. (preview)
- Authentication: https://hips.com/docs/reference/authentication : API keys, Authorization: Bearer, key kinds and the workspace states that refuse a key. (preview)
- Chat completions: https://hips.com/docs/reference/chat-completions : POST /v1/chat/completions: the request, the model field, the response, streaming, tool calls and structured outputs. (preview)
- Responses: https://hips.com/docs/reference/responses : POST /v1/responses, served after validation of its compatibility rows. (preview)
- Models: https://hips.com/docs/reference/models : GET /v1/models and GET /v1/models/{id}: the list shape with the hips entry per model and deployment. (preview)
- Tasks: https://hips.com/docs/reference/tasks : POST /v1/tasks: the synchronous task envelope, batching, per-input errors and the usage object. (preview)
- Jobs: https://hips.com/docs/reference/jobs : POST /v1/jobs, GET /v1/jobs/{id}, GET /v1/jobs and POST /v1/jobs/{id}/cancel: states, cancellation and result retrieval. (preview)
- Uploads: https://hips.com/docs/reference/uploads : POST /v1/uploads: a presigned PUT for a task or job input; the purpose, statuses, mismatch and retention. (preview)
- Deployments: https://hips.com/docs/reference/deployments : The read side of customer model deployments: dep_ ids in the model field and in deployment, states, lanes. (preview)
- Requirements and headers: https://hips.com/docs/reference/headers : The four request requirements as headers and as the hips body object, every x-hips-* response header, the x-ratelimit-* headers. (preview)
- Errors: https://hips.com/docs/reference/errors : One envelope on every public route, the closed list of types, and every code of this version. (preview)
- Usage: https://hips.com/docs/reference/usage : Usage fields per billing unit; money separate from units; the usage record behind every request. (preview)
- Versioning: https://hips.com/docs/reference/versioning : How the contract is versioned and deprecated: the path carries the version, models are versioned separately, unknown is an error. (preview)
- Operating guidance: https://hips.com/docs/operating : Limits, errors, retries and billing outcomes, idempotency, streaming, data handling, and the compatibility matrix. (preview)
- Limits: https://hips.com/docs/operating/limits : Rate limits per key and plan, the x-ratelimit-* headers, what 429 means. No number until decided. (preview)
- Errors: https://hips.com/docs/operating/errors : Every error type with when it occurs and what to do; only 503 and 500 are ever 5xx; 503 carries Retry-After. (preview)
- Retries and billing outcomes: https://hips.com/docs/operating/retries : What is charged when an attempt fails, is interrupted, times out or is retried. Placeholders until decided. (preview)
- Idempotency: https://hips.com/docs/operating/idempotency : The idempotency-key header on the task and job routes; what a repeat answers. (preview)
- Streaming: https://hips.com/docs/operating/streaming : Server-sent events on chat completions, cancellation, what a canceled stream reports, where the cost of a streamed answer arrives. (preview)
- Data handling: https://hips.com/docs/operating/data : Weights versus request data, the import path for customer models, the request-data policy by plan, pools and runtime access, regions. (preview)
- Compatibility matrix: https://hips.com/docs/operating/compatibility : Status per endpoint, per model and per client version: supported, partial, unsupported, unvalidated. Every cell unvalidated at launch. (preview)
- Docs changelog: https://hips.com/docs/changelog : Dated entries, newest first, each naming the pages changed. The docs are versioned by date; the API is versioned by its path. (preview)