kepta

Python client for KEPTA — local memory for AI agents.

Your agents forget you after every conversation. KEPTA fixes that — with a SQLite file on your own machine. No account, no cloud, no telemetry.

This package is the client, not the app. KEPTA runs as a desktop application on the same device; here you connect to it.

pip install kepta

In thirty seconds

from kepta import KeptaClient

kepta = KeptaClient()          # finds the running instance on its own

kepta.save("Carbonara", "Guanciale, pecorino, egg yolk. No cream.", tags=["cooking"])

for hit in kepta.search("carbonara without cream"):
    print(f"{hit.score:.2f}  {hit.memory.title}")

How search works

Three tracks, fused by Reciprocal Rank Fusion: full text (BM25), vectors and the knowledge graph.

Full text and graph work immediately. The vector track also finds what is worded differently from the question — for that it needs a local embedding model:

ollama pull nomic-embed-text

Only then does search("what do I cook with pasta") find the carbonara recipe, which does not contain the word pasta at all. Without the model, hit.vector_score stays at 0.0 and search remains lexical — not an error, just fewer hits on paraphrased questions. How to tell:

kepta.health()["embeddings"]     # {'total': 128, 'embedded': 128, ...} — or zeros everywhere

Why this is interesting

The same memory as your agents. Claude Desktop and Cursor talk to the same database over MCP. What your Python script writes, Claude knows in its next answer.

Memories age. Each one has a type, a validity window and a confidence score. When someone moves house, the new address supersedes the old one — the old one stays as history and drops in the ranking. Contradictions do not pile up.

old = kepta.save("Home", "Alex lives in Hamburg.")
kepta.update(old.id, valid_to=1788400000000)      # expired from this point on
kepta.save("Home, current", "Alex now lives in Leipzig.")

m = kepta.list()[0]
m.is_expired, m.is_superseded                      # state right on the object

No dependencies. Standard library only. A memory that promises privacy should not pull foreign code into your process.

Finding the connection

KeptaClient() looks in this order:

1. Environment variable KEPTA_URL
2. ~/.kepta/endpoint.json — the address file KEPTA writes on startup
3. http://127.0.0.1:3000 as a fallback for development mode

Step 2 is the important one: the packaged app picks a random port. Being explicit works too, of course:

kepta = KeptaClient("http://127.0.0.1:52341")

If nothing is running you do not get a cryptic network error but a sentence that says what to do:

if not kepta.is_alive():
    print("KEPTA is not running — start the app or set KEPTA_URL.")

What the client can do

Memory and SearchHit are frozen dataclasses with type annotations. Alongside the overall score, SearchHit exposes the individual tracks as vector_score and lexical_score.

The server behind it

The client talks to the KEPTA HTTP API on your own machine — start it from the open core (npm run dev), or run KEPTA Pro, the desktop app for macOS and Windows with the knowledge graph, drag & drop import and a chat cockpit. Every download starts with one free Pro day, then it runs free with daily limits and never locks; a license (€12/month or €120/year) removes the limits (write to me on LinkedIn — or reach the buy button inside the app). German notes work as well as English ones: the search stopword list covers both languages.

License

MIT