MemTetherBLOCK
Cross-client AI memory hub — one physical SQLite, many clients. File-level pointers, REST API, Web Dashboard, MCP. Zero cloud. Dual-timeline governance. PyPI: memtether
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
Your agent remembers things. Who audits what it remembers?
MemTether is a local-first memory hub for multi-client AI agents. While most memory systems compete on retrieval accuracy, MemTether competes on governance: who wrote each fact, when it took effect, what it superseded, which conflicts a human adjudicated, and a tamper-evident hash chain anyone can verify — no code, no dependencies. Retrieval is provided (SGM deterministic SQL routing + hybrid embedding), but it is not the headline: agent memory that you can audit, correct, and trust is.
Coming from mem0, Zep, or Letta? examples/mem0_migration.py shows a complete round-trip migration from a mem0 JSON export — no mem0 SDK required.
$ memtether court fact-20261007-abc +============================================================+ | MEMTETHER MEMORY COURT - CASE FILE | | 1. CONTENT Clash proxy port is 7897 | | 2. TEMPORAL valid: 10-05 → present | known: 10-05 → now | | 3. PROVENANCE #1 remember by=codex → #2 correct by=user | | 4. CONFLICTS case #3: vs fact-...old verdict=superseded | | 5. INTEGRITY chain: PASS (3 anchors, 42 entries hashed) | +============================================================+
TL;DR
- Your AI clients forget each other? Every agent keeps its own memory. MemTether gives them one shared, governed brain via file-level pointers — no cloud, no sync.
- Your AI writes things it should not? Every memory carries a source identity and a hash-chained audit trail. Tamper with the database and the chain detects it — run
python examples/audit_chain_demo.pyto see it turn red. - Your AI changed its mind? Corrections never delete. Bi-temporal supersession keeps who-changed-what-when-why forever, queryable at
6cc5008e1812OBSERVED · 2026-10-11Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add memtether -- None memtether==0.1.0a73
Trust audit
BLOCKgrade F · trust 52/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | WARN |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | FAIL |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (1 observation(s))
- Network
- declared (5 observation(s))
- Shell
- declared (7 observation(s))
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (24)
yaml.load(text, Loader=_Loader)
<!DOCTYPE html>
eaf_artifact_bundle.zip
mod = __import__(os.path.splitext(os.path.basename(cred_path))[0])
__import__(m)
__import__(_m)
mod = __import__(os.path.splitext(os.path.basename(cred_path))[0])
__import__(m)
# -*- coding: utf-8 -*-
.pre-commit-config.yaml
.release-baseline.json
.mcp.json.template
h = hashlib.md5(text.encode('utf-8')).hexdigest()[:12]rid = 'miss-%s' % hashlib.md5(query.encode('utf-8')).hexdigest()[:10]hashlib.md5(('%f' % time.time()).encode()).hexdigest()[:8])h = hashlib.md5()
h = hashlib.md5((c + "test").encode()).hexdigest()[:12]
if curl -sf http://127.0.0.1:8820/health > /dev/null; then break; fi
curl -sf http://127.0.0.1:8820/health
BODY=$(curl -sf http://127.0.0.1:8820/dashboard)
"openai_base_url": "http://127.0.0.1:7874/v1",
client = OpenAI(api_key=WB2API_KEY, base_url='http://127.0.0.1:7874/v1')
for ci, conv in enumerate(locomо if False else locomo):
Gates applied: no_behavioural_pass.
6cc5008e1812full audit observations/trust-audit/mcp-server/memtether__memtether.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-11 | 6cc5008e1812 | BLOCK | F | 52 | first audit |
Questions
What is the MemTether MCP server?
Cross-client AI memory hub — one physical SQLite, many clients. File-level pointers, REST API, Web Dashboard, MCP. Zero cloud. Dual-timeline governance. PyPI: memtether
Is MemTether safe to connect to an agent?
No — not without reading the findings first. The audit graded it F (52/100) and found 2 critical or high issues in the source. Each one is listed on this page with the file and line it is on.
What credentials does MemTether need?
It reads DEEPSEEK_OFFICIAL_KEY, GPTX_ASTRA_KEY, MEMTETHER_API_KEY, MEM_EXTRACT_API_KEY and ZHIPU_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does MemTether run?
It speaks stdio, so it runs as a local process your client starts. It is published on PyPI as memtether.
How current is this page?
The grade is for one exact copy of the source (6cc5008e1812), read on 2026-10-11. The repository is watched and re-audited when it changes.