agent-memory

Nestor gives AI systems durable memory through a local CLI, HTTP API, and Docker stack. The primary interface is the nestor CLI: use it to write facts, retrieve relevant memories, reinforce useful memories, connect context, inspect diagnostics, and evaluate simple production rules.

What You Can Do

• Store typed memories as chunks with slots.
• Retrieve memories by cues and context, with score diagnostics.
• Record practice so repeated use changes retrieval strength.
• Associate chunks so current context can spread activation.
• Set buffers for goal, retrieval, imaginal, and task state.
• Evaluate production rules against memory and buffers.
• Run all of this locally through the CLI, HTTP API, or Docker stack.

Start With The CLI

When Nestor is installed, use the CLI directly:

From this repository, run the same commands through Cargo:

Start with nestor guide, then open the command group or workflow you need:

Full CLI docs live in docs/cli/README.md.

Common Workflows

Start or point at a Nestor API before running commands that call the service. For local development:

In another terminal, check connectivity:

Reinforce it and connect it to context:

Set a buffer and inspect metrics:

For complete copy-pasteable workflows, see docs/cli/workflows.md.

Run Nestor Locally

Use the in-memory local API when you want the fastest CLI loop:

Use Docker Compose when you want the local API, Memgraph, and Prometheus stack: