NewsCatcher launches free NewsMCP for AI agents

NewsCatcher has launched NewsMCP, a free news MCP server that gives AI agents clustered story events instead of long lists of duplicate article links. The service is live now with no signup required and includes tools for search, story expansion and coverage checks.

Why it matters: - AI agents often waste context on dozens of near-identical links about the same news event. - NewsMCP is designed to return one clustered record per story, which can help agents answer current-events questions with less noise. - Each event includes an independent-newsroom count, giving agents a built-in signal for how widely a claim has been reported.

What happened: - NewsCatcher launched NewsMCP on September 7, 2026. - NewsMCP is a news MCP server built for AI agents and is free to use with no account required. - The service is available now at newsmcp.com, with a browser-based playground, one-step setup for Claude Code and Cursor, and a REST interface for agents not using MCP.

The details: - NewsMCP returns live news as deduplicated story events rather than article links. - The service clusters coverage from multiple outlets into a single event with citations. - Article body text is never returned; agents follow a source link when they need the full piece. - NewsMCP includes five read-only tools: news for searching story events, get_story for expanding one event, check_coverage for measuring how widely a claim has been reported, and check_limits and check_health for diagnostics. - The keyless tier allows 20 calls per hour, a seven-day lookback and 20 stories per call. - A free API key raises usage to 50 calls per hour, a fourteen-day lookback and 50 stories per call. - Commercial use is permitted on every tier, including the keyless tier.

Between the lines: - NewsCatcher is positioning NewsMCP as a separate product, not a wrapper around its existing News API. - Co-founder Artem Bugara said APIs and MCP servers solve different problems because an API serves a pipeline, while an MCP server serves a language model inside a live conversation. - NewsMCP is built around the context-window limits of LLMs, which makes deduplication and source-count signals more useful than bulk article text. - The product runs on the same news infrastructure that serves NewsCatcher's enterprise customers, which suggests the company is extending existing data assets into a new AI-native interface.

What's next: - NewsCatcher says NewsMCP is open for immediate use through the website, the playground and integrations for developer tools. - Agents that need broader access can use the free API key tier, while commercial users can start on any tier. - NewsCatcher is likely betting that AI-native retrieval will become a separate layer from traditional news APIs, especially for current-events workflows.

The bottom line: - NewsCatcher is trying to make news retrieval more usable for AI agents by collapsing duplicate coverage into one event, not many links.