| Aspect | Engram (engram.ai) | Mem0 | Zep |
|--------|---------------------|------|-----|
| Starting Price | Not publicly verifiable | Public pricing page available (includes a free tier) | Public pricing page available |
| Recommended for | Unclear due to limited public information | Teams seeking a packaged memory layer with managed options | Organizations seeking temporal graph-based memory |
| Key Strength | Ambiguous product identity / limited public signals | Broad mindshare in "agent memory layer" category | Temporal knowledge graph approach |
| Key Weakness | Ambiguous product identity; limited reviews | Requires tuning; potential for costs to scale with usage | Added architectural complexity; published pricing may be heavier than DIY |
- Direct competitors and their key differentiators
- Mem0: Positions as an AI memory layer with explicit pricing and enterprise claims; verify current security posture with primary documentation.
- Zep: Offers a temporal knowledge graph approach and a public pricing page.
- LangGraph Platform / LangChain ecosystem: Provides a strong orchestration ecosystem; memory is often implemented as part of workflow/state management rather than as a dedicated memory product.
- DIY stack (Postgres + pgvector + BM25/hybrid search): Lower vendor lock-in but requires engineering investment.
- Pricing comparison: value proposition vs alternatives
- For many teams, the key comparison is:
- Paying for a managed memory layer versus building on Postgres/pgvector
- Paying for a vector DB service versus keeping everything in Postgres
- Customer testimonials comparing products
- No robust, third-party evidence of customers switching from a named competitor to engram.ai was found.
- Some vendor benchmark pages explicitly compare retrieval to Mem0 and LangMem/LangGraph, but these are vendor-published comparisons rather than third-party customer testimony.
- Market share and adoption trends
- There is insufficient reliable data available for engram.ai to estimate market share or adoption trends.