# Pearl Inference Platform - [Build on Pearl Inference Platform](https://pearl-research.mintlify.dev/docs/index.md): OpenAI-compatible inference supporting the latest open-weight models, get started in minutes - [Quickstart](https://pearl-research.mintlify.dev/docs/quickstart.md): Make your first request to Pearl Inference in a few minutes. - [Models](https://pearl-research.mintlify.dev/docs/models.md): The catalog is deliberately curated: production models spanning reasoning, coding, chat, and vision, all behind the same OpenAI-compatible endpoint. - [Concepts](https://pearl-research.mintlify.dev/docs/concepts.md): The handful of ideas behind the platform: organizations, API keys, prepaid credits, and one OpenAI-compatible endpoint. - [Data residency](https://pearl-research.mintlify.dev/docs/data-residency.md): Control which geographic region processes your inference traffic, for compliance, latency, or data-handling requirements. - [Model deprecation](https://pearl-research.mintlify.dev/docs/deprecation.md): Plan your migration with deprecation dates and recommended replacement models. - [Overview](https://pearl-research.mintlify.dev/docs/inference.md): One OpenAI-compatible endpoint: what requests look like, what comes back, and how to handle the failure modes. - [Streaming](https://pearl-research.mintlify.dev/docs/streaming.md): Receive the response token by token as it is generated, instead of waiting for the full reply. - [Function calling](https://pearl-research.mintlify.dev/docs/function-calling.md): Let the model ask your code to run a function, then continue with its result, the primitive behind every agent. - [Structured outputs (JSON mode)](https://pearl-research.mintlify.dev/docs/json-mode.md): Constrain the response to syntactically valid JSON at decode time, then validate the shape against your schema. - [Reasoning models](https://pearl-research.mintlify.dev/docs/reasoning.md): Every model in the catalog can think before it answers, and shows you the thinking in a separate field. - [Vision](https://pearl-research.mintlify.dev/docs/vision.md): Ask questions about images with GLM-5.3 Flash, sent inline as base64 data URIs. - [Prompt caching](https://pearl-research.mintlify.dev/docs/prompt-caching.md): The network reuses computation for prompt prefixes it has recently seen, automatically, with no code changes. - [Plans & rate limits](https://pearl-research.mintlify.dev/docs/plans.md): Start on Basic with lower limits, unlock higher Basic limits with a credit purchase of at least $5, or request Pro or Enterprise. - [Credits & payments](https://pearl-research.mintlify.dev/docs/credits.md): Inference draws down a prepaid USD balance. Top it up when you need to, there are no subscriptions or minimum commitments. - [Auto-recharge](https://pearl-research.mintlify.dev/docs/auto-recharge.md): When your balance dips below a threshold you set, we charge your saved card and top it back up, no one in the loop. - [Usage & analytics](https://pearl-research.mintlify.dev/docs/analytics.md): The Analytics tab reads from your billing data, what you see is exactly what you're billed. - [Team & roles](https://pearl-research.mintlify.dev/docs/team.md): One organization, one balance, one catalog, with roles scoping who can spend, ship, and administer. - [API keys](https://pearl-research.mintlify.dev/docs/api-keys.md): Organization-scoped bearer keys, shown once at creation, everything after that is lifecycle management. - [Connect your tools to Pearl](https://pearl-research.mintlify.dev/docs/integrations.md): Use your existing coding agent or SDK with a Pearl endpoint, API key, and model ID. - [OpenCode setup](https://pearl-research.mintlify.dev/docs/integrations/opencode.md): Connect your terminal coding agent with a custom Pearl provider. - [Claude Code setup](https://pearl-research.mintlify.dev/docs/integrations/claude-code.md): Run Claude Code with a Pearl model through the Messages API. - [Codex setup](https://pearl-research.mintlify.dev/docs/integrations/codex.md): Run the Codex CLI with a Pearl model through the Responses API. - [Pi setup](https://pearl-research.mintlify.dev/docs/integrations/pi.md): Connect Pi to Pearl with a custom model provider. - [Open WebUI setup](https://pearl-research.mintlify.dev/docs/integrations/open-webui.md): Chat with Pearl models in your own Open WebUI instance. - [n8n setup](https://pearl-research.mintlify.dev/docs/integrations/n8n.md): Build workflows and agents using a Pearl chat model. - [LiteLLM setup](https://pearl-research.mintlify.dev/docs/integrations/litellm.md): Call Pearl models through LiteLLM's OpenAI-compatible Python interface. - [Cline setup](https://pearl-research.mintlify.dev/docs/integrations/cline.md): Use Pearl from the OpenAI Compatible provider in the Cline extension for VS Code. - [Zed setup](https://pearl-research.mintlify.dev/docs/integrations/zed.md): Use a Pearl model in the Zed editor and its agent. - [Goose setup](https://pearl-research.mintlify.dev/docs/integrations/goose.md): Connect the Goose agent to Pearl with a custom provider. - [Aider setup](https://pearl-research.mintlify.dev/docs/integrations/aider.md): Edit code with Aider using a Pearl model. - [Oh My Pi setup](https://pearl-research.mintlify.dev/docs/integrations/oh-my-pi.md): Use Pearl as a custom provider in Oh My Pi. - [OpenAI SDK setup](https://pearl-research.mintlify.dev/docs/integrations/openai-sdk.md): Use Chat Completions with the OpenAI Python SDK. - [OpenAI Responses setup](https://pearl-research.mintlify.dev/docs/integrations/openai-responses.md): Use stateless Responses with the OpenAI Python SDK. - [Anthropic SDK setup](https://pearl-research.mintlify.dev/docs/integrations/anthropic-sdk.md): Use Messages and token counting with the Anthropic Python SDK. - [Coding agent setup](https://pearl-research.mintlify.dev/docs/guides/coding-agents.md): Any agent that speaks the OpenAI protocol can run on Pearl, three settings, no code changes. - [Migrate from OpenAI](https://pearl-research.mintlify.dev/docs/guides/migrate-from-openai.md): Same SDK, same request shapes, the migration is two constructor arguments, a model-name map, and a short list of differences to check. - [Overview](https://pearl-research.mintlify.dev/docs/reference/inference.md): The data-plane API your applications call to run models: OpenAI- and Anthropic-compatible endpoints behind one base URL. - [Chat completions](https://pearl-research.mintlify.dev/docs/reference/chat-completions.md): The primary inference endpoint, OpenAI-compatible chat with streaming, tools, JSON mode, reasoning, and vision. - [Responses](https://pearl-research.mintlify.dev/docs/reference/responses.md): Stateless response creation with native JSON and SSE responses. - [Messages](https://pearl-research.mintlify.dev/docs/reference/messages.md): Native Anthropic Messages responses and SSE events for compatible models. - [Count tokens](https://pearl-research.mintlify.dev/docs/reference/messages/count-tokens.md): Count a Messages request's input tokens without generating. - [Completions (legacy)](https://pearl-research.mintlify.dev/docs/reference/completions.md): Raw-prompt text completion, the pre-chat OpenAI shape, served for tooling that still speaks it. - [List models](https://pearl-research.mintlify.dev/docs/reference/models.md): The live catalog, one entry per served model, with context windows, capabilities, and supported parameters. - [Overview](https://pearl-research.mintlify.dev/docs/reference/management.md): Read your organization's credits and usage from scripts, dashboards, and finance tools, and let any API key check its own limit and spend. - [Rate limits](https://pearl-research.mintlify.dev/docs/reference/management/rate-limits.md): Every endpoint allows a fixed number of requests per minute, for each organization or, on GET /key, for each key. - [Get credits](https://pearl-research.mintlify.dev/docs/reference/management/credits.md): Every credit your organization has been granted and how much of it has been used. `total_credits` minus `total_usage` is what remains. Figures can trail recent usage by a few minutes. - [Get current API key](https://pearl-research.mintlify.dev/docs/reference/management/key.md): Describes the key that calls it: its name, credit limit and when it resets, expiration, and spend. Any API key can call it, inference keys included, so an application can check its own key's remaining budget. The `usage_` fields come from the same usage history as the Analytics page, so they can tra… - [Get activity](https://pearl-research.mintlify.dev/docs/reference/management/activity.md): Usage per UTC day and model over the last 30 completed days, oldest first. Today is not included until it ends. A day or model without usage has no row. ## OpenAPI Specs - [inference](/docs/openapi/inference.yaml) - [management](/docs/openapi/management.yaml) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.