LLM routing, without the glue code
What is LLM routing on Backboard?
Backboard lets you call 17,000+ models from a single endpoint and change which model you use at any time, without rewriting your app. Instead of hard‑coding every provider's SDK and payload quirks, you:
LLM ROUTING
Why engineers route through Backboard
From benchmark-leading memory to BYOK with no token markup — everything built into one stateful API.
One API, 17,000+ models, many free
Call OpenAI, Anthropic, Google Gemini, Cohere, xAI, open‑source models via OpenRouter, and more through a single integration. Tap into many free and low‑cost models for experimentation and background workloads.
BYOK with no token markup
Use your own API keys for OpenAI, Anthropic, Google Gemini, Cohere, xAI, OpenRouter, and others. Backboard does not add a premium on tokens—your model costs are whatever your provider charges.
Stateful by default
Backboard is built around state. You don't resend full history every call; portable state and memory travel with the conversation.
Adaptive context management built in
Switch between models with different context windows and Backboard automatically fits the conversation. It reserves ~20% of context for raw state and key elements and frees the rest via intelligent summarization, prioritizing system prompts, recent turns, tool calls, RAG, and search.
Configurable memory, RAG, and tools on every route
Attach long‑term memory (lite or pro), RAG, and web search to any model through the same endpoint. You keep one orchestration surface; models become pluggable.
how it works
How LLM routing works
You call a single msg‑style endpoint and pass the model (or routing rule), the state or conversation ID, and optional tools: memory, RAG, web search, custom tools.
routing patterns
Routing patterns you can implement
Same state, same memory, same tools—different models for different jobs.
how it works
Why not just build your own router?
Wiring a couple of models is easy. The hard parts are what Backboard solves.
Wiring a couple of models is easy. The hard parts:
Keeping state and memory consistent across models and providers
Handling different context windows without losing important info
Tracking cost, latency, and usage when logic is scattered
Making RAG, web search, and tools behave the same for every model
Managing multiple keys and pricing models without accidentally overpaying
Backboard gives you:
A unified API for 17,000+ models
BYOK support for major providers with no token markup
Access to many free models for experimentation and background work
Free state management and Adaptive Context Management baked in
Best‑in‑class configurable memory (lite and pro), plus RAG and web search
You integrate once and get world‑leading routing everywhere.