# Impala AI Documentation

- [Introduction](https://docs.getimpala.ai/introduction.md): The most intelligence per dollar on any open-source model, running on your reserved throughput or in your own cloud.
- [Run async and open source](https://docs.getimpala.ai/async-inference.md): Cheaper inference, better quality. The best open-source models cost far less to run when your workload can tolerate latency.
- [Run in our cloud, or yours](https://docs.getimpala.ai/deployment-options.md): Impala runs in our cloud or inside your own AWS account.
- [Quickstart](https://docs.getimpala.ai/impala-cloud-setup.md): Get your endpoint and key, then send your first request.
- [Chat completions](https://docs.getimpala.ai/send-a-request.md): The request shape, the parameters that matter for async work, and what changes from OpenAI.
- [Rate limit webhooks](https://docs.getimpala.ai/rate-limit-webhooks.md): Get an event when your team's RPM or TPM limit changes, so your router can follow it.
- [Run a batch](https://docs.getimpala.ai/run-your-first-batch.md): This guide shows how to submit batch inference to your Impala endpoint using the OpenAI-compatible API. You’ll upload an input file, then create a batch.
- [Jobs](https://docs.getimpala.ai/jobs-and-job-ids.md): A job is a model plus a configuration, provisioned for your account. Passing a job ID is optional; without one, Impala picks the job for you.
- [Images](https://docs.getimpala.ai/pass_images_with_image_url.md): Send images to batch chat-completion requests by S3 reference instead of inline base64.
- [Video](https://docs.getimpala.ai/video-input.md): Send video to vision-language models in batch chat-completion requests.
- [Impala Swarm](https://docs.getimpala.ai/swarm.md): Run an agent swarm on Impala — the harness, the fan-out and the model calls in one place, next to your data.
- [Impala Herd](https://docs.getimpala.ai/impala-herd.md): Herd is the inference engine every Impala workload runs on. It reads the shape of your live traffic and reconfigures itself around it — kernels, batching, decoding.
- [LoRA adapters](https://docs.getimpala.ai/lora-adapters.md): Serve your fine-tuned model on Impala. Bring a LoRA adapter, and we deploy it for you to call through serverless or batch.
- [Impala Leap](https://docs.getimpala.ai/weight-sync.md): Impala Leap swaps model weights in seconds, with no redeploy. Upload a checkpoint, reload it, and the running fleet serves it.
- [Install Impala in your VPC](https://docs.getimpala.ai/getting-started-infrastructure-setup.md): Deploy the Impala data plane in your AWS account.
- [Expose the OpenAI API endpoint](https://docs.getimpala.ai/exposing_the_openai_api_endpoint_service.md): Make your Impala OpenAI API proxy reachable to internal teams and workloads over AWS PrivateLink.
- [Raise AWS service quotas](https://docs.getimpala.ai/aws_service_quotas_for_Impala_data_plane_deployment.md): Review and raise the AWS service quotas required before deploying the Impala data plane with ./deploy.sh.
- [LiteLLM](https://docs.getimpala.ai/connect-litellm.md): Point your LiteLLM gateway at your Impala endpoint.
- [Anthropic SDKs and Claude Code](https://docs.getimpala.ai/anthropic-sdk.md): Impala serves the Anthropic Messages API. Point the Anthropic SDKs, the Claude Agent SDK, or Claude Code at your endpoint.
- [Security & data handling](https://docs.getimpala.ai/security.md): Impala runs single-tenant, in our cloud or in yours. Your data stays isolated and isn't stored.
- [Further reading](https://docs.getimpala.ai/further-reading.md): Impala's engineering writing on inference economics, kernels and expert parallelism.
- [Glossary](https://docs.getimpala.ai/glossary.md): The vocabulary these docs use for async inference, agent workloads and Impala's own objects.
- [Health](https://docs.getimpala.ai/api-reference/system/health.md): Liveness probe - returns 200 if server is running.
- [Ready](https://docs.getimpala.ai/api-reference/system/ready.md): Readiness probe - returns 200 if services are initialized and ready.
- [List Files](https://docs.getimpala.ai/api-reference/files/list-files.md): List all uploaded files.
- [Upload File](https://docs.getimpala.ai/api-reference/files/upload-file.md): Upload a file using true streaming without buffering to disk.
- [Get File](https://docs.getimpala.ai/api-reference/files/get-file.md): Retrieve metadata for a specific file.
- [Delete File](https://docs.getimpala.ai/api-reference/files/delete-file.md): Delete a specific file.
- [Get File Content](https://docs.getimpala.ai/api-reference/files/get-file-content.md): Retrieve the content of a specific file.
- [List Batches](https://docs.getimpala.ai/api-reference/batches/list-batches.md): List all batches.
- [Create Batch](https://docs.getimpala.ai/api-reference/batches/create-batch.md): Create a new batch of API requests.
- [Get Batch](https://docs.getimpala.ai/api-reference/batches/get-batch.md): Retrieve a specific batch.
- [Cancel Batch](https://docs.getimpala.ai/api-reference/batches/cancel-batch.md): Cancel a batch.

## OpenAPI Specs

- [openapi](/api-reference/openapi.json)
- [openai-api-swagger](/openai-api-swagger.json)
- [openapi](/openapi.json)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.