Chroma MCP vs MCP Fetch Server

A side-by-side look at pricing, capabilities, pros, cons, and our editorial scores.

	Chroma MCP MCP Servers	MCP Fetch Server MCP Servers
Tagline	Official MCP server that gives LLM clients direct access to the Chroma vector database.	Official Model Context Protocol reference server that lets any MCP-compatible LLM fetch web pages and read them as clean markdown.
Category	MCP Servers	MCP Servers
Pricing	Free· Open source (Apache 2.0). Free to run locally or self-hosted; embedding-function API keys (OpenAI, Cohere, Jina, VoyageAI, Roboflow) billed by those providers. Chroma Cloud pricing set separately by Chroma.	Free· Free and open source (MIT License). No hosted tier; you run it locally via uvx, pip, or Docker.
Model	—	—
Editorial score	—	—
Use cases	Long-term memory for Claude DesktopTeam knowledge base shared across agentsRAG over local documentsSemantic code search in Cursor or ContinueVector store for custom MCP agentsPersonal notes and journal recallMetadata-filtered document retrievalChroma Cloud access from LLM clients	Reading documentation pages inside Claude DesktopGrabbing a GitHub README before scaffolding a projectSummarising a long article for a chat agentChecking a package changelog during code reviewAnswering questions from a linked spec section by sectionFeeding a URL into an MCP-driven research agentPulling clean markdown for a note-taking workflowVerifying a citation URL inside an editorial pipeline
Pros	Official server from the Chroma team, tracks the database's features closely Four deployment modes (ephemeral, persistent, HTTP, Chroma Cloud) from one binary Twelve well-scoped MCP tools cover the full collection and document lifecycle Supports six embedding functions including OpenAI, Cohere, Jina, and VoyageAI with per-collection persistence Zero-config install via uvx - drops into Claude Desktop or Cursor in a couple of JSON lines Apache 2.0 licensed and readable Python source; easy to fork or extend Query tools expose HNSW tuning and metadata/full-text filters, not just naive semantic search	Official reference implementation maintained in the modelcontextprotocol/servers repo — tracks the spec and won't rot behind third-party churn. Zero-install path via `uvx mcp-server-fetch`; Docker image and pip package also available for sandboxed or reproducible setups. HTML-to-markdown conversion produces context-window-friendly output instead of raw tag soup. `max_length` + `start_index` let a model chunk through long pages without blowing its context. Optional Node.js fallback swaps in a more robust HTML simplifier when present. First-class configuration for User-Agent, robots.txt obedience, and outbound proxy — practical knobs for real-world scraping etiquette. Works out of the box with every major MCP client (Claude Desktop, Claude Code, VS Code, Cursor, Zed, Continue).
Cons	You still have to run and back up the Chroma store yourself unless you pay for Chroma Cloud Command-line API key arguments are convenient but leak into process lists; env/.env path is safer but extra setup No built-in access control or multi-tenant isolation - anything the MCP client sees, it can delete Embedding-function persistence only works for collections created on Chroma v1.0.0+; older stores need migration Python-only server; teams on pure Node stacks add a runtime dependency MCP tool surface is CRUD-shaped - no higher-level RAG primitives like chunking, re-ranking, or hybrid fusion	No JavaScript rendering — SPAs, Cloudflare-challenged pages, and content loaded after DOMContentLoaded come back empty or as a shell. No built-in anti-bot handling; sites that block plain HTTP clients (many news sites, LinkedIn, X) will 403. Security caveat called out in the README: the server can reach local/internal IPs, so it is a genuine SSRF risk if exposed to an untrusted model or user. Single tool, single verb — for crawling, sitemap walks, or extracting structured data you need a heavier server like Firecrawl-MCP or a Playwright-based one. Truncation defaults to 5,000 characters; models that don't understand `start_index` can silently miss content below the fold.
Website	github.com	github.com

Pick Chroma MCP if

✅ Official server from the Chroma team, tracks the database's features closely
✅ Four deployment modes (ephemeral, persistent, HTTP, Chroma Cloud) from one binary
✅ Twelve well-scoped MCP tools cover the full collection and document lifecycle
✅ Supports six embedding functions including OpenAI, Cohere, Jina, and VoyageAI with per-collection persistence

Pick MCP Fetch Server if

✅ Official reference implementation maintained in the modelcontextprotocol/servers repo — tracks the spec and won't rot behind third-party churn.
✅ Zero-install path via `uvx mcp-server-fetch`; Docker image and pip package also available for sandboxed or reproducible setups.
✅ HTML-to-markdown conversion produces context-window-friendly output instead of raw tag soup.
✅ `max_length` + `start_index` let a model chunk through long pages without blowing its context.

Compare a different pair →