Best Letta (MemGPT) MCP Server Alternatives 2026

Knowledge graph-based persistent memory system. Store and retrieve contextual information.

Free and open-source by Anthropic. Knowledge graph stored locally — no external service costs.

Dynamic and reflective problem-solving through thought sequences.

Free and open-source by Anthropic. Runs locally with no external API dependencies.

Exa's official MCP server connects AI assistants to a search engine purpose-built for AI, using neural embeddings to match on meaning rather than keywords so agents get clean, ready-to-use content instead of a page of blue links to re-parse. The default tool set covers web_search_exa for quick topical lookups and web_search_advanced_exa for full control over domains, date ranges, and content filters, plus specialized tools for code_search (searching real-world code and GitHub), company_research (building company profiles, competitor lists, and financials), crawling/web_fetch (pulling clean content from a specific URL), people_search and linkedin_search (public professional-profile lookups), and deep_researcher_start/check for long-running multi-step research tasks backed by Exa's Research API. The server is hosted at https://mcp.exa.ai/mcp — no local process to run — and connects via one-line setup in Cursor, VS Code, Claude Code, Claude Desktop (available as a native Connector), Codex, OpenCode, Windsurf, and Antigravity, authenticated with an EXA_API_KEY from the Exa dashboard. Tool exposure is tunable per client via a ?tools= query parameter on the endpoint URL, letting teams ship narrow, purpose-built configurations (e.g. company-research-only or LinkedIn-only agents) instead of exposing the full surface, and Exa ships ready-made Claude Skills/agent definitions for common patterns like company research and people search with built-in query-variation and token-isolation guidance.

The MCP server is free and open-source. Exa: Free tier with 1,000 searches/mo. Pro plans available for higher volume. See official pricing.

Neo4j graph database server (schema + read/write-cypher) and graph database backed memory.

The MCP server is free and open-source. Neo4j Aura: Free tier available. Professional: from $65/mo. Enterprise: Custom pricing. Community edition is free.

Search, Query and interact with data in your Milvus Vector Database.

Free and open-source MCP server. Milvus is open-source. Zilliz Cloud (managed): Free tier available, paid plans for production use.

Embeddings, vector search, document storage, and full-text search with the open-source AI application database.

Free and open-source MCP server. Chroma is open-source and free to self-host. Chroma Cloud currently in development.

OpenAI does not publish a dedicated, first-party "MCP server" for its own API — a `openai/mcp-server` repo does not exist. Instead, OpenAI's official open-source contribution to the MCP ecosystem is on the client side: openai/openai-agents-python (27,000+ stars), a lightweight framework for building multi-agent workflows with the OpenAI API that ships native support for connecting to MCP servers as a tool source, letting an OpenAI-model-powered agent call out to any MCP server (filesystem, GitHub, databases, etc.) the same way a Claude-based agent would. In other words, OpenAI's MCP investment is "consume MCP tools from an OpenAI agent," not "expose OpenAI itself as an MCP server." Teams that specifically want to call OpenAI's chat, embeddings, or image-generation endpoints as MCP tools from Claude, Cursor, or another MCP client instead rely on small community-built wrapper servers around the OpenAI SDK, authenticated with an `OPENAI_API_KEY`, exposing tools like generate_completion, generate_embedding, or generate_image. Typical use of the Agents SDK side: build a Python agent that uses GPT models for reasoning while pulling live context through an MCP filesystem or web-search server. Update this entry if OpenAI ships a genuine first-party MCP server for its own API in the future.

The MCP server is free and open-source. OpenAI API: Pay-per-token pricing. GPT-4o: ~$5/1M input tokens. GPT-3.5 Turbo: ~$0.50/1M input tokens. See official pricing.

Connect to Hugging Face Hub APIs - search spaces, papers, explore datasets and models.

The MCP server is free and open-source. Hugging Face: Free tier for public models. Pro: $9/mo. Enterprise Hub: Custom pricing. Inference API has free and paid tiers.

Open-source tool for collaborative editing, versioning, evaluating, and releasing prompts.

The MCP server is free and open-source. Langfuse Cloud: Free tier (50K observations/mo). Pro: from $59/mo. Enterprise: Custom. Self-hosted is free.

The Atlassian Remote MCP Server brings Confluence and Jira into any MCP-compatible AI assistant, IDE, or agent platform through a centrally hosted, enterprise-grade connection backed by Atlassian's Teamwork Graph. Launched in May 2025 with Anthropic as the first official partner and hosted on Cloudflare infrastructure, authentication is handled via OAuth 2.1 — no local server process to deploy or maintain. For Confluence specifically, available operations include summarizing pages and spaces, creating new pages from AI-generated content, searching across your wiki with natural language, and performing multi-step knowledge retrieval across Confluence spaces. Jira operations include creating, updating, and triaging work items, summarizing sprint state, and linking knowledge to in-flight issues. Atlassian's Teamwork Graph underpins every response — connecting people, services, knowledge, and work items into a unified context for richer AI answers. Enterprise customers at AT&T, NVIDIA, Pfizer, Booking.com, and Visa use the integration in production. Connect from Claude Desktop via Settings > Connectors, or from Claude Code with: `claude mcp add --transport http atlassian https://mcp.atlassian.com/v1/mcp`. Cursor and Windsurf users can add the remote URL directly to their MCP config.

The MCP server is free and open-source. Confluence: Free tier (up to 10 users). Standard: $6.05/user/mo. Premium: $11.55/user/mo. Enterprise: Custom.