Brave Search (official)
BraveBrave's own server, and the one to use now that the older reference implementation sits in the archived repository
Brave Search (official) vs Tavily MCP
Brave's own server, and the one to use now that the older reference implementation sits in the archived repository On the other side: web search built for agents rather than people: results come back with extracted content and optional answer synthesis, which removes a second fetch step
Installnpx -y @brave/brave-search-mcp-servernpx -y tavily-mcp
SourceOfficialOfficial
Deciding factorThe agent should cite sources a human can open and check.versus: The agent needs to read several pages quickly and report a conclusion.
Bottom line
Brave when citations matter; Tavily when reading several pages quickly matters.
The short answer
Brave returns search results — titles, URLs, snippets — and lets the agent decide what to fetch. Tavily returns extracted content and an optional synthesis, which removes a step and adds an opinion. For citation work Brave is more honest; for research summaries Tavily is faster.
Facts
| Option | Install | Source | Link |
|---|---|---|---|
| Brave Search (official) | npx -y @brave/brave-search-mcp-server | Official | Repository ↗ |
| Tavily MCP | npx -y tavily-mcp | Official | Repository ↗ |
Pick Brave Search (official) if
The agent should cite sources a human can open and check.
Pick Tavily MCP if
The agent needs to read several pages quickly and report a conclusion.
What would change our mind
A project going archived, or a maintainer changing the install path, flips this comparison faster than any feature release. Both entries are re-checked weekly, and the date above tells you how much to trust the details.
Read the full entries
More head-to-heads
Filesystem vs Desktop Commander
Gives an agent scoped read, write, and search access to directories you explicitly allow
Comparison
Git vs GitHub
Read-only access to a repository's history: status, diffs, commit logs, and blame
Comparison
Playwright MCP vs Browser MCP
Drives a real browser through accessibility snapshots rather than screenshots, so an agent can click, type, and assert without a vision model
Comparison
Memory vs AgentDB
A persistent knowledge graph stored in a local JSON file, exposed through entities, relations, and observations
Comparison
Kubernetes MCP vs Kubernetes MCP (kubectl)
Talks to the Kubernetes API directly instead of shelling out to kubectl, so an agent can list contexts, read pod logs, and apply manifests without parsing command output
Comparison
LaunchDarkly MCP vs Unleash MCP
Reads and changes feature flags through the LaunchDarkly API, so an agent can turn a flag off during an incident instead of asking a human to open a dashboard
Comparison