Your First Memory

Push your first piece of knowledge to Memax and understand how it gets processed.

Push a document

The simplest way to add knowledge is to push a file:

memax push --file ./ARCHITECTURE.md

✓ Saved — "Architecture Overview" (core · private) [52ms]

Memax automatically:

Extracts content from the file
Classifies it into a category (e.g., core, decisions, reference)
Chunks it semantically for efficient retrieval
Embeds each chunk as a vector for similarity search
Assigns a boundary (private by default)

More ways to push

From stdin

echo "We chose PostgreSQL over MongoDB because we need strong consistency for memory deduplication" | memax push

With metadata

memax push --file ./runbook.md \
  --category reference \
  --tags "ops,incidents,runbook" \
  --hint "on-call runbook for production incidents"

URLs

memax push https://docs.google.com/document/d/1abc...

Memax fetches, summarizes, and stores a link reference. Use --snapshot to also capture the full content at that point in time.

Multiple files

memax push -f ./docs/*.md

Sync an entire directory

memax sync ./docs

This watches for changes and keeps your hub in sync with the directory.

Recall it back

Now test retrieval:

memax recall "how does the architecture work?"

─── Architecture Overview (core · private) ──────────────────
The system uses a three-layer architecture: API server (Go),
web app (Next.js), and background worker (River queue)...

Relevance: 0.94 · Pushed 2 minutes ago

The recall command performs semantic search — it finds content by meaning, not just keyword matching.

View your memories

# List recent memories
memax list

# Show a specific memory
memax show mem_a1b2c3d4

The real power comes when agents recall context automatically. Set up an agent integration next.

On this page