For Agent Developers¶

Build memory into your own agents using the memsearch CLI and Python API.

Install¶

# pip
pip install memsearch

# or uv (recommended)
uv add memsearch

Optional embedding providers

pip install "memsearch[onnx]"    # Local ONNX (recommended, no API key)
# or: uv add "memsearch[onnx]"

# Other options: [openai], [google], [voyage], [ollama], [local], [all]

Quick Example¶

from memsearch import MemSearch

mem = MemSearch(paths=["./memory"])

# Index markdown files
await mem.index()

# Search
results = await mem.search("Redis config", top_k=3)
for r in results:
    print(r["heading"], r["score"])

CLI¶

memsearch index ./memory                    # index markdown files
memsearch search "batch size" --top-k 5     # semantic search
memsearch expand <chunk_hash>               # expand a chunk
memsearch watch ./memory                    # live file watcher

See the full CLI Reference → and Python API →.

How Plugins Use the API¶

All 4 platform plugins are built on top of the same CLI/API:

Plugin Capture:  conversation → LLM summary → append daily.md → memsearch index
Plugin Recall:   memsearch search → memsearch expand → parse-transcript

If you're building a plugin for a new platform, see the Architecture and existing plugin source code in plugins/.