Profile-based long-term memory backend that extracts and consolidates conversation turns into structured user profile slots (topics, preferences, facts) stored in PostgreSQL. Incoming data is queued in a Redis buffer and batch-processed by an LLM into the profile, decoupling write cost from request latency. At read time the pre-compiled profile is returned via SQL lookup (sub-100ms); an event timeline of timestamped entries sits alongside the profile and supports full-text and tag-based search. Self-hostable via Docker Compose (FastAPI + Postgres + Redis); managed cloud also available.
- Storage
- PostgreSQL (structured user profile slots + event timeline) with a Redis buffer for incoming unprocessed data; cloud-managed or self-hosted via Docker Compose
- Retrieval
- SQL lookup of pre-compiled profile slots at sub-100ms latency; no per-query LLM call at retrieval time; optional embedding search over the event timeline for semantic context; context assembled into a prompt prefix on demand
- Self-host
- Self-host: moderate
- License
- Apache-2.0
- Pricing
- Self-hosted OSS (Apache-2.0) is free and remains installable from the repo. The managed cloud is effectively gone: memobase.io and memobase.io/pricing both return HTTP 404 and the hosted playground at app.memobase.io returns a Cloudflare 522, re-confirmed on 2026-07-23 and again on 2026-08-01 — so no paid tier can be purchased and none is quoted here. The previously carded 'free tier plus pay-as-you-go' description is withdrawn rather than restated, since nothing substantiates it now. · Free / OSS
- GitHub stars
- 2,835
- Last release
- 2026-01-11
- Last commit
- 2026-01-11
- First catalogued
- 2026-06-28
Strengths
- Sub-100ms profile retrieval: pre-compiled slots are SQL-read at inference time, eliminating per-query LLM calls
- Batch buffer amortizes LLM extraction cost; self-reported ~5x cheaper and ~5x faster than mem0 at equivalent quality
- Multi-language SDK surface: Python, Node.js/TypeScript, and Go SDKs plus a REST API and a stdio MCP server
- Event timeline alongside the structured profile enables temporal queries, full-text search, and tag-based filtering over past interactions
- Controllable profile schema: operators define slot types and token budgets, giving predictable cost and output shape
Watch out
- Profile-centric design targets user-preference and fact memory; episodic or agent-task memory is outside the primary scope
- No formal GitHub Releases — version history is git tags only; latest tag v0.0.42 on 2026-01-11 (no commits since, ~6.5 months as of 2026-07-23); verify maintenance status before adopting
- Performance claims (5x cheaper/faster vs mem0, LOCOMO SOTA) are self-reported without published backbone/embedder configs; no independent reproductions available
- As of 2026-07-23 the marketing site (memobase.io) 404s and the hosted playground (app.memobase.io) is unreachable (Cloudflare 522); the PyPI package also lags the repo, last published v0.0.27 in Nov 2025 vs the v0.0.42 git tag. The same org (memodb-io) has shipped active new development on a separate project, Acontext, since — evaluate self-hosting only and confirm the hosted/managed path is still live before committing to it
- Marked deprecated on 2026-08-01. Every signal checked a second time, five weeks after the first: homepage and /pricing still 404, hosted playground still 522, still zero commits since 2026-01-11 (~202 days), PyPI still stuck at v0.0.27 from 2025-11-08, and open community PRs from June 2026 still unmerged. The vendor has published no sunset notice, so this is a verdict from evidence, not an announcement. The Apache-2.0 code is still there and still runs — but nobody is maintaining it, and the managed product it was sold with no longer exists
Best for
- Chat applications and AI companions requiring persistent, structured user profiles (preferences, topics, facts) at sub-100ms retrieval latency
- Multi-session educational tutors, customer-service bots, or voice agents (LiveKit adapter available) where long-term user history must be structured and queryable
- Teams wanting a self-hostable or cloud-managed memory REST backend with multi-language SDKs and an MCP server for IDE or agent integration
How it integrates
Benchmark results
No sourced results yet.
Sources
- Vendor homepage — HTTP 404 again on 2026-08-01, five weeks after the same 404 was recorded on 2026-07-23; memobase.io/pricing also 404s and app.memobase.io/playground still returns Cloudflare 522. Basis for moving status needs-review → deprecated and priceTier free-paid → free (third-party)
- PyPI JSON API re-check — latest published version still 0.0.27, uploaded 2025-11-08 (~9 months stale, and well behind the repo's v0.0.42 git tag); corroborates that the project is no longer shipping (third-party)
- Monthly card re-verification — GitHub API repo metadata: 2,812 stars, license Apache-2.0 confirmed, archived=false, last commit on the default branch 2026-01-11 (third-party)
- https://www.memobase.io/ (vendor)
- https://github.com/memodb-io/memobase (vendor)
- https://github.com/memodb-io/memobase/tags (vendor)
- https://docs.memobase.io/llms.txt (vendor)
- https://docs.memobase.io/cost (vendor)
- https://www.pulsemcp.com/servers/memobase (third-party)
- Monthly re-check — homepage returns HTTP 404 (confirmed on repeated requests); repo homepage field still points here (third-party)
- Hosted playground app — returns HTTP 522 (Cloudflare: origin connection timed out) (third-party)
- GitHub API — Apache-2.0 license confirmed, not archived, pushed_at still 2026-01-11 (no new commits) (third-party)
- PyPI JSON API — latest published version 0.0.27, uploaded 2025-11-08, well behind the repo's v0.0.42 git tag (third-party)
- GitHub API for memodb-io/Acontext, a separate project from the same org ('Context Data Platform that Improves your Agent with Experiences', linked from the memobase README) — 3584 stars, pushed 2026-07-14, suggesting the org's active development has shifted here (third-party)
Last verified 2026-08-01 · updated by refresh-framework-cards