TenseMem — vendor claims and pricing, checked_on 2026-10-29 (UTC+8)
Sources: https://github.com/tensemem/tensemem (README at commit 48a9cf7), https://pypi.org/project/tensemem/
Repo: 7 stars, Apache-2.0, last push 2026-10-25 19:25 UTC+8. PyPI: 0.1.0 uploaded 2026-10-21 23:12 UTC+8 (earlier 0.0.1).
Pricing: free, open source (Apache-2.0), pip install tensemem. No hosted service or account. Optional LLM extractors use your own OpenAI/Anthropic/Gemini/Ollama/LiteLLM access.
README claims (verbatim):
- "One current truth per ONE slot. A new city supersedes the old one. History still answers "where did I live before?""
- "Every read reports freshness. as_of_seq, pending, is_current. Wait for a message, or see that it is not applied yet."
- "No self-echo. Recalled context() is fenced. Extract needs a verbatim quote from the user."
- "Corrections stick. correct() restores the displaced value and survives rebuild() with a newer model."
- "The engine decides. The LLM emits typed ops. Deterministic rules apply them."
- "No infrastructure. Default store is one SQLite file inside your process."
- "`tensemem bench` scores the packaged 101 conversations with FakeExtractor (no network)": TenseMem 1.00 / 0.00 / 1.00 / 0.00 (current-state accuracy / stale-fact leak / freshness honesty / contamination).
- Quickstart: mem = Memory(llm=FakeExtractor()); mem.register_predicate("lives_in"); mem.add(...); mem.context(user_id=..., min_seq=..., wait=1.5)