SleepSweet
SleepSweet is an open-source memory hygiene and audit layer for AI agents: an
offline engine plus local CLI and MCP server that deduplicates, resolves
conflicts, flags stale facts, finds broken links and dead paths in the memory
your agent already keeps — CLAUDE.md, AGENTS.md, ~/.claude auto-memory,
Cursor rules, and GitHub Copilot instructions. It is not a memory store and
it does not do retrieval.
pip install sleepsweet # or zero-install: uvx sleepsweet run
sleepsweet run # scan + report + one confirmation + fixes with a log
sleepsweet run --llm claude # + an LLM judge: fact conflicts, confirmed merges
sleepsweet run --yes # apply fixes without asking (hooks, automation)
sleepsweet report # the latest report; --details for every finding
sleepsweet diff # what changed in memory since the previous run
sleepsweet undo # revert applied fixes; --to DATE rolls back to a date
sleepsweet asof 2026-08-01 # what the agent knew at that date
Everything deterministic — duplicates (semantic ones included: the embedding
model ships inside the package), stale facts, dead paths, broken links,
trust scoring — works with no LLM at all. Plugging in a judge adds fact
conflict detection and merge confirmation, and takes one string:
--llm claude (or claude:<model>), --llm ollama:<model>,
--llm command:<any CLI>, or --llm https://host/v1#<model> — same via the
SLEEPSWEET_LLM environment variable. Your model, your key; prompts are
redacted before they leave the process.
Local-first: no telemetry, no network calls, secrets redacted before
anything is stored. Non-destructive: nothing changes without your y,
history is marked, never deleted, and every decision is journaled — the
event log can reconstruct what your agent knew at any past date.
Agents can run SleepSweet themselves via MCP (sleepsweet mcp): scan, read
the stable JSON report, judge fact conflicts, dismiss false positives — no
API key or model configuration needed.
Docs and source: https://github.com/sleepsweet/sleepsweet
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file sleepsweet-0.3.1.tar.gz.
File metadata
- Download URL: sleepsweet-0.3.1.tar.gz
- Upload date:
- Size: 28.5 MB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/7.0.0 CPython/3.11.4
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
9eb5ac578927ec7385ebaafcbd0a675c2ae6f202e22085fd5c64d67c36251dee
|
|
| MD5 |
f38251410ff5b4e89f6098a17969f491
|
|
| BLAKE2b-256 |
d17bfc5ae4a89cb0be5020786187d1024c53da24a494e477def9f4a6e54f8dbd
|
File details
Details for the file sleepsweet-0.3.1-py3-none-any.whl.
File metadata
- Download URL: sleepsweet-0.3.1-py3-none-any.whl
- Upload date:
- Size: 28.4 MB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/7.0.0 CPython/3.11.4
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
11e9932372220d84547ac5ba8782126c840554e4a25478835c9c2eeb8f10ead7
|
|
| MD5 |
5fcea3d5c2c15250d417f266c436e5b9
|
|
| BLAKE2b-256 |
c30b9b93efec627875839d4a35a29182a01f0003e76cdef34ebea0236e3a1775
|