Tiger Memory MCP Server

Official MCP server from Timescale for Tiger — a memory and context management system built on PostgreSQL and pgvector. Enables AI agents to persist, retrieve, and manage long-term memory through semantic vector search — providing durable agent memory backed by Timescale's PostgreSQL infrastructure.

Evaluated Mar 06, 2026 (0d ago) vcurrent
Homepage ↗ Repo ↗ Databases timescale memory vector postgresql official tiger mcp-server pgvector
⚙ Agent Friendliness
78
/ 100
Can an agent use this?
🔒 Security
86
/ 100
Is it safe for agents?
⚡ Reliability
77
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
78
Documentation
78
Error Messages
76
Auth Simplicity
80
Rate Limits
78

🔒 Security

TLS Enforcement
95
Auth Strength
88
Scope Granularity
82
Dep. Hygiene
80
Secret Handling
85

Official Timescale. HTTPS. Database credentials. pgvector memory may contain sensitive agent data. SOC2.

⚡ Reliability

Uptime/SLA
82
Version Stability
76
Breaking Changes
74
Error Recovery
76
AF Security Reliability

Best When

A team building production AI agents needs durable memory with semantic retrieval — backed by Timescale's PostgreSQL infrastructure with pgvector for vector similarity search.

Avoid When

You use other vector databases (Pinecone, Weaviate, Qdrant) for memory. Timescale Tiger is PostgreSQL-centric — evaluate if your infrastructure aligns.

Use Cases

  • Persistent agent memory across conversations using semantic vector search
  • Storing and retrieving structured knowledge from long-running AI agent workflows
  • Building agent memory systems on PostgreSQL/pgvector infrastructure
  • Managing AI context windows with selective memory retrieval from Timescale

Not For

  • Teams not using Timescale/PostgreSQL infrastructure
  • Simple session-only memory without persistence needs
  • Non-AI agent use cases without semantic search requirements

Interface

REST API
Yes
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
Yes
Webhooks
No

Authentication

Methods: username_password api_key
OAuth: No Scopes: No

Timescale service credentials required. PostgreSQL connection string with pgvector. Timescale Cloud or self-hosted PostgreSQL.

Pricing

Model: freemium
Free tier: Yes
Requires CC: No

Timescale Cloud has a free trial. Self-hosted PostgreSQL with pgvector is fully free. MCP server is free open source from official Timescale org.

Agent Metadata

Pagination
cursor
Idempotent
Full
Retry Guidance
Documented

Known Gotchas

  • pgvector embedding dimensions must match your embedding model — configure before storing memories
  • Memory retrieval quality depends on embedding model quality — use consistent models
  • Connection pooling important for agents making frequent memory queries
  • Official from Timescale — high quality but Tiger is relatively new compared to other Timescale products

Alternatives

Full Evaluation Report

Detailed scoring breakdown, competitive positioning, security analysis, and improvement recommendations for Tiger Memory MCP Server.

$99

Scores are editorial opinions as of 2026-03-06.

5220
Packages Evaluated
26151
Need Evaluation
173
Need Re-evaluation
Community Powered