{"id":"hosted-agent-memory","slug":"hosted-agent-memory","name":"Hosted agent memory layer","shortName":"Hosted agent memory layer","problemStatement":"Bolt persistent memory (add/retrieve) onto any agent without building a vector stack.","whenNotTopPick":"Need temporal knowledge graphs — Zep; stateful agent runtime — Letta; already have pgvector.","categoryIds":["cat-vector-memory"],"depotPickProviderId":"mem0","depotPickRationale":"DX + pricePerf for bolt-on memory; watch retrieval metering.","runnersUpIds":["zep","letta","supabase-pgvector","supermemory","memobase","cognee"],"weights":{"capabilityFit":0.26,"pricePerformance":0.12,"latencyReliability":0.12,"developerDx":0.22,"privacyPosture":0.12,"lockInRisk":0.08,"maturity":0.08},"methodologyNotes":"DX + pricePerf for bolt-on memory; watch retrieval metering.","tags":["hosted"],"depth":{"whenNotToUse":{"text":"Need temporal knowledge graphs — Zep; stateful agent runtime — Letta; already have pgvector.","source":"editorial `whenNotTopPick` (data/jtbds.json)"},"whenNotToUseReason":null,"runnerUp":{"providerId":"zep","name":"Zep","overallForJob":7.48,"pickOverallForJob":7.63,"beatsPickOn":["capabilityFit","latencyReliability","privacyPosture"],"source":"first resolvable entry of the job's editorially ranked `runnersUpIds`; scores are stored job-weighted scores"},"runnerUpReason":null,"lockIn":{"pickLockInRisk":5,"runnerUpLockInRisk":5.2,"scale":"0–10 stored curated estimate; higher = more lock-in (worse)","pickExitSignals":["open source (provider tag `oss`)"],"runnerUpExitSignals":["open source (provider tag `oss`)"],"pickNotes":[{"text":"self-host OSS if caps hurt.","field":"scores.notes"}],"selfHostFallback":[],"migration":"The pick has lower stored lock-in risk than runner-up Zep (5 vs 5.2). Pick exit signals: open source (provider tag `oss`).","migrationReason":null,"derivedFrom":"stored lockInRisk scores, provider tags (oss, self-host, openai-compat, open-models), free rows that list this job, and verbatim provider notes. No migration guide was invented."}},"freshness":{"pick":{"score":{"asOf":"2026-09-14","ageDays":43,"stale":true,"staleAfterDays":30,"reason":null},"lastReviewed":{"asOf":"2026-09-14","ageDays":43,"stale":true,"staleAfterDays":30,"reason":null},"sourceCheck":{"checkedAt":"2026-10-27","sourceUrls":3,"sourcesReachable":2,"pricingRows":0,"pricingRowsReconfirmed":0,"note":"A source re-check confirms the page is reachable and, for pricing rows, that the stored amount and SKU name appear on it. It does not re-derive or change the score.","reason":null}}},"depotPick":{"id":"mem0","slug":"mem0","name":"Mem0","tagline":"Default hosted agent memory bolt-on. Model retrieval volume carefully; self-host OSS if caps hurt.","overall":7.63},"runnersUp":[{"id":"zep","slug":"zep","name":"Zep","tagline":"Pick when agents need temporal/entity graph memory, not just vector snippets. Mem0 for simpler recall.","overall":7.48},{"id":"letta","slug":"letta","name":"Letta Cloud","tagline":"Choose when the agent IS a long-lived stateful runtime. For bolt-on memory use Mem0/Zep.","overall":7.41},{"id":"supabase-pgvector","slug":"supabase-pgvector","name":"Supabase (pgvector)","tagline":"Depot default for most agent apps under ~25–50M vectors. Don't add Pinecone until Postgres actually hurts.","overall":8.63},{"id":"supermemory","slug":"supermemory","name":"Supermemory","tagline":"Managed context engine when you want memory without running a vector DB. Mem0 for simpler fact API; Cognee…","overall":7.94},{"id":"memobase","slug":"memobase","name":"Memobase","tagline":"Depot pick for portable/user-profile memory (MCP across Claude/ChatGPT/agents). Mem0 remains general…","overall":7.95},{"id":"cognee","slug":"cognee","name":"Cognee","tagline":"Graph memory when documents→ontology matter. Zep for temporal bipartite graphs; Mem0 for simplest hosted…","overall":8.1}]}