# Aquifer > A self-hosted load balancer for agent and API traffic: it stores excess work durably, paces it into the target API at the rate the target asks for, keeps tenants fair, and returns results synchronously, over a WebSocket, or by webhook. Each page below is plain Markdown. Fetch only the ones you need; https://rjpruitt16.github.io/aquifer/llms-full.txt has all of them in one file. ## Start here - [Why Aquifer](https://rjpruitt16.github.io/aquifer/overview.md): What Aquifer is, the problem it solves, and where it sits next to a gateway, Canalis and Inlet. - [Quick start](https://rjpruitt16.github.io/aquifer/quick-start.md): Run Aquifer from the binary, Docker Compose, Fly.io, AWS, or a Canalis fleet. - [How it works](https://rjpruitt16.github.io/aquifer/how-it-works.md): The job lifecycle and the delivery guarantee (at least once). - [Use cases](https://rjpruitt16.github.io/aquifer/use-cases.md): MCP tools, protecting an HTTP API, webhook intake, outbound pacing, the /proxy reverse proxy, WebSockets and cross-region redirect. ## Guides - [Dynamic pacing](https://rjpruitt16.github.io/aquifer/pacing.md): How the target API sets the pace with X-Aqueduct-* headers: per-tenant queues, fair admission, sizing the backlog budget, slow start and the ORCA fallback. - [Streaming](https://rjpruitt16.github.io/aquifer/streaming.md): Stay on the line over SSE or /proxy, or hang up and get a webhook; when the webhook is skipped. - [Pools](https://rjpruitt16.github.io/aquifer/pools.md): Load balancing across registered members, weighted by declared capacity and live reputation. - [Partitioning](https://rjpruitt16.github.io/aquifer/partitioning.md): Static partitioning, regional cluster routing, drain mode and external registration. - [Deployment](https://rjpruitt16.github.io/aquifer/deployment.md): Sidecar, standalone, embedded or behind a Gateway API proxy; choosing a machine size. - [Integrations](https://rjpruitt16.github.io/aquifer/integrations.md): Inlet, for provider webhook storms (GitHub, Shopify, Stripe, Anthropic). - [L8 protocol](https://rjpruitt16.github.io/aquifer/l8.md): Webhook delivery signed with public keys, optional payload encryption and request schemas. - [Reliability](https://rjpruitt16.github.io/aquifer/reliability.md): The durable queue, crash and panic recovery, and job TTLs. ## Reference - [API overview](https://rjpruitt16.github.io/aquifer/api-overview.md): The POST /jobs request and the warning about untrusted callers. - [API reference](https://rjpruitt16.github.io/aquifer/api.md): Every endpoint: jobs, /proxy, streams, WebSockets, health, metrics, webhooks, admission and cluster routing. - [Configuration](https://rjpruitt16.github.io/aquifer/configuration.md): Rate limits per target API in YAML, and the environment variables. - [Configuration rationale](https://rjpruitt16.github.io/aquifer/configuration-rationale.md): Why the admission defaults are what they are. - [Framework adapters](https://rjpruitt16.github.io/aquifer/adapters.md): The HTTP, MCP stdio and A2A front doors. - [Adapters reference](https://rjpruitt16.github.io/aquifer/adapters-reference.md): The MCP tools, the A2A Agent Card, and writing your own adapter, storage backend or metrics adapter. - [Drain mode](https://rjpruitt16.github.io/aquifer/drain-mode.md): Handing an instance to one tenant at a time: the state machine, settings and ledger payloads. - [External registration](https://rjpruitt16.github.io/aquifer/registration.md): Reporting an instance to an external control plane: settings and payloads. - [Benchmarks](https://rjpruitt16.github.io/aquifer/benchmarks.md): Throughput, crash recovery, memory, capacity per machine size, and the GPU run. ## Optional - [Source code](https://github.com/rjpruitt16/aquifer) - [Aquifer pilot skill](https://github.com/rjpruitt16/aquifer/blob/main/skills/aquifer-pilot/SKILL.md): evaluates Aquifer against your own workload in a shadow run - [Aquifer engineer skill](https://github.com/rjpruitt16/aquifer/blob/main/skills/aquifer-engineer/SKILL.md): benchmarks, tunes, partitions and scales an Aquifer deployment (with or without Inlet) - [L8 specification](https://rjpruitt16.github.io/l8-protocol/)