1
Neon's MCP server: manage serverless Postgres with natural language
★ 649 · +-1 this week · MIT · updated 2026-09-30
8.0/10
What it does Neon MCP is Neon's official server for managing serverless Postgres from a chat or coding agent. An agent can create a project, open a branch, design tables, run SQL and check slow queries. Its standout feature is safe migrations. The agent first applies a schema change to a temporary branch, where you can test it, and only… Read more →
Pros
- Branch-based safe migrations
- Free tier
Pricing: Freemium · postgres database
🔥 Deal: Startups: up to $200K (VC-backed) in credits
2
Official MongoDB and Atlas MCP server
★ 1.1k · +1 this week · Apache-2.0 · updated 2026-10-01
7.7/10
What it does MongoDB MCP Server is MongoDB's official server for connecting AI agents to MongoDB databases and Atlas, MongoDB's cloud service. With a connection string, the agent can explore databases and collections, infer a collection's schema, run finds and aggregations, check indexes and query plans, and, if you allow it, insert, update or delete data. With Atlas service-account credentials,… Read more →
Pricing: Free · mongodb database
🔥 Deal: MongoDB for Startups: tiered Atlas credits plus Voyage AI tokens for companies under 7 years old, Series A or earlier
3
Official Qdrant server that gives agents a semantic memory: store and search text in a vector DB
★ 1.5k · +3 this week · Apache-2.0 · updated 2026-09-04
7.6/10
What it does mcp-server-qdrant is Qdrant's official MCP server. It gives an agent a semantic memory layer on top of the Qdrant vector database. It is deliberately small: the agent stores a piece of text with optional metadata, the server embeds it with a local FastEmbed model and writes it to a collection, and later the agent retrieves the closest… Read more →
Pros
- Official Qdrant server, simple to set up with uvx or Docker
- Local embeddings via FastEmbed, works with local, self-hosted or cloud Qdrant
- Read-only mode and customisable tool descriptions
Cons
- Only two tools; no update, delete or list operations
- Only FastEmbed embedding models are supported
Pricing: Open source · qdrant vector-database semantic-search memory embeddings
🔥 Deal: Qdrant for Startups: 20% off Qdrant Cloud for 12 months (pre-seed/seed/Series A, under 5 years old, <$5M raised)
4
Open-source context database giving agents shared memory, resources and skills
★ 39k · +450 this week · AGPL-3.0 · updated 2026-10-02
7.5/10
What it does OpenViking, from Volcengine, is an open-source context database for AI agents. It gives an agent one place to keep project knowledge, user memories and reusable skills across sessions. Everything is organised as a virtual filesystem under viking:// URIs, so an agent browses, reads, writes and searches context the same way it works with files. What is inside… Read more →
Pros
- Filesystem-style viking:// context with tiered summaries
- Hooks plus MCP integrations for Claude Code, Codex, Cursor and more
- Very active development with SDKs in three languages
Cons
- Needs an embedding model and a VLM plus a running server
- AGPL terms matter if you offer it as a service
Pricing: Open source · agent-memory context-database rag mcp-server self-hosted
5
Read-only-by-default MySQL access with opt-in writes, SSH tunnels and PII masking
★ 2.1k · +1 this week · MIT · updated 2026-07-27
7.2/10
What it does MCP Server for MySQL connects a coding agent to a MySQL database so it can read the schema and run SQL while it builds or debugs your app. A typical session has the agent inspect tables before writing a migration or model, check what data a failing endpoint actually returns, or explain a slow query. The server… Read more →
Pros
- Read-only by default, with writes enabled per operation and per schema
- Built-in SSH tunnel, multi-database mode and optional PII redaction
Cons
- A single generic mysql_query tool, with no dedicated schema or explain tools
- Community project, not from Oracle or MySQL
Pricing: Open source · mysql sql database read-only ssh-tunnel