Aider has no plugin system and no memory
The proxy path fixes that.
No MCP support to configure. Point one environment variable at MemoryRouter and every chat is remembered.
One environment variable. Permanently.
14 days free
Full access. Point Aider at the proxy, watch it remember, cancel anytime.
See the difference
You: Finish the migration script we discussed
Aider: I don't have any prior context on a migration script. What is the target schema?
A new terminal session, and yesterday's conversation is gone.
You: Finish the migration script we discussed
Aider: Continuing the users table split we planned, Postgres 16, zero-downtime constraint. Writing the backfill script now.
No tool call needed. It was already in context.
Sound familiar?
Re-explaining the same architecture in every new terminal session
No plugin ecosystem to add memory through
Decisions from yesterday's session, gone
With MemoryRouter
Every request through the proxy is stored automatically, no tool calls needed
Relevant memory is retrieved and injected before every prompt reaches the model
Works with any OpenAI-compatible provider Aider already supports
14-day free trial. Cancel anytime.
Connect Aider in two minutes
No plugin, no MCP config. Just point Aider's OpenAI-compatible settings at MemoryRouter.
export OPENAI_API_BASE=https://api.memoryrouter.ai/v1
export OPENAI_API_KEY=mk_your_memory_keyOn Windows: setx OPENAI_API_BASE https://api.memoryrouter.ai/v1 && setx OPENAI_API_KEY mk_your_memory_key, then restart your shell.
$ aider --model openai/gpt-5.5Replace mk_your_memory_key with your MemoryRouter Memory Key. See the full supported model list, including Anthropic, Google, and OpenRouter models, in the API docs.
Sign up, copy your key, paste it in. Two minutes.
How it works
Without MemoryRouter:
Aider Model provider Response (gone when the session ends)
With MemoryRouter:
Aider MemoryRouter proxy (retrieve, inject, store) Model provider Response
MemoryRouter is an OpenAI-compatible endpoint that sits in the request path. It queries your memory, injects relevant context into the prompt, forwards the request to your chosen provider using the provider key on your account, stores new memory, and returns the standard response. Aider never knows the difference.
No plugin, no MCP
Two environment variables. Aider does not need any code changes or extensions to work with MemoryRouter.
Every model Aider supports
OpenAI, Anthropic, Google, Meta, Mistral, xAI, DeepSeek, and 200+ models through OpenRouter, all with the matching provider prefix.
Toggle memory per key
Suffix your key with :read, :write, or :off to change behavior for one session without touching any config.
Same memory everywhere else
One vault. Connect it to Claude, ChatGPT, Cursor, or Claude Code and the memory follows you between tools.
Connect in two minutes. 14-day free trial.
Start free. Then just $20/mo.
14-day free trial. Cancel anytime.
- 200M memory tokens / mo included
- Unlimited projects
- Same memory in ChatGPT, Claude, Claude Code, and Codex
- 300 to 500 ms average recall on a typical vault
- Your keys, your data, your control
No inference markup, ever.
FAQ
Give Aider a memory tonight
14 days free. Two minutes to connect. Tomorrow's first session already knows everything from today.