Memory in plain SQL — no vector DB, fully inspectable, portable. LoCoMo benchmark: 81.95% accuracy at ~1,294 tokens/query (self-reported). Paper: arxiv.org/abs/2603.19935.
- Storage
- Any SQLite / Postgres / MySQL; BYODB option supported for self-hosted databases
- Retrieval
- Recalls memories via plain SQL queries over relational tables — no vector database required; works with any LLM + datastore combination.
- Self-host
- Self-host: trivial
- License
- Apache-2.0
- Pricing
- Open source: free (self-hosted, BYODB). Memori Cloud: Free (5,000 memories created/mo, 15,000 recalled/mo); Team from $60K/yr (single production agent, managed multi-tenant cloud); Business from $150K/yr (multiple production agents, single-tenant managed cloud); Enterprise custom (org-wide unlimited agents, customer-VPC or on-prem). Paid tiers include unlimited memories, pooling, access control, SSO/SCIM. · Free + paid
- GitHub stars
- 15,744
- Last release
- 2026-05-28
- Last commit
- 2026-07-28
- First catalogued
- 2026-06-28
Strengths
- SQL-queryable / fully debuggable memory
- ~1,294 tokens/query on LoCoMo (self-reported) — ~5% of full-context cost
- One-line SDK integration: .llm.register(client)
- BYODB: bring your own SQLite / Postgres / MySQL
- Captures agent actions (tool calls, decisions) not just conversation text
- First-party plugins for Claude Code, OpenClaw, and Hermes, plus a remote-HTTP MCP server that needs no SDK or code changes
Watch out
- Simpler retrieval than graph/hybrid systems — may miss relational or temporal queries
- Self-reported benchmark (81.95% LoCoMo); backbone LLM not stated in source
- Memori Cloud's paid production tiers are enterprise-priced (Team from $60K/yr, Business from $150K/yr, as of Jul 2026) — there is no self-serve monthly plan between the free cloud tier and a five/six-figure annual contract; cost-sensitive teams needing paid cloud memory at small scale should evaluate the open-source BYODB path instead
Best for
- Cost-sensitive production: skip the vector DB and run on the SQL infra you already have
- Inspectable, debuggable memory you can query directly
- Agents where what the agent *did* (tool calls, decisions) matters as much as what the user *said*
How it integrates
Benchmark results
| Benchmark | Value | Backbone | Trust | Source |
|---|---|---|---|---|
| locomo | 81.95 accuracy | — | Self-reported | Memori (MemoriLabs) ↗ |
Sources
- Pricing page re-checked — unchanged: Cloud Free $0 (5,000 memories created, 15,000 recalled); Team from $60K/yr (single production agent, managed multi-tenant); Business from $150K/yr (multiple production agents, single-tenant); Enterprise custom private offer (customer-VPC/on-prem) (vendor)
- Monthly card re-verification — GitHub API repo metadata: 15,674 stars, GitHub reports no auto-detectable SPDX license (card license recorded from the repo's own LICENSE file), archived=false, last commit on the default branch 2026-07-28 (third-party)
- Memori repo (GitHub now resolves GibsonAI/Memori → MemoriLabs/Memori) (vendor)
- MemoriLabs — product home (vendor)
- Memori paper (MemoriLabs) (paper)
- GitHub API license endpoint returns 'Other'/NOASSERTION (detection quirk); decoded LICENSE file content confirms standard Apache License 2.0 text — Apache-2.0 value retained (third-party)
- GitHub API re-check — not archived, pushed 2026-06-15T23:18Z (last commit); latest release v3.3.6 (2026-05-28), unchanged since prior verification (third-party)
- Memori pricing page — replaces the prior placeholder 'managed cloud Free/Paid' value with concrete tiers: Cloud Free (5,000 memories created/mo, 15,000 recalled/mo); Team from $60K/yr; Business from $150K/yr; Enterprise custom (vendor)
- Memori MCP server docs — stateless HTTP transport at api.memorilabs.ai/mcp/, header-based auth, works with any MCP client (vendor)
- memori on PyPI — v3.3.6, Apache-2.0 classifier (vendor)
- @memorilabs/memori on npm registry — latest 0.0.11 (beta 0.1.25-beta) (third-party)
- Memori repo integrations/ directory — confirms first-party claude-code, hermes, and openclaw plugin subdirectories exist with README/plugin manifests (third-party)
- Memori Claude Code plugin docs (vendor)
- Memori OpenClaw plugin quickstart docs (vendor)
- Memori Hermes plugin docs (vendor)
- Memori CLI quickstart docs — `python -m memori sign-up <email>` for account/rate-limit management (vendor)
Last verified 2026-08-01 · updated by refresh-framework-cards