Unbiased AI-powered news
Most news APIs serve raw content with every outlet’s framing baked in. Substrate runs each story through a multi-source synthesis pipeline first — extracting claims, stripping editorial spin, and scoring confidence — so an agent can read one article instead of ten.
We’re building this. The pipeline, scoring, and transparency layer that the API would expose are already running in production on the web — the remaining work is the public endpoint, auth, quotas, and documentation. No pricing, launch date, or SLA yet; we’d rather under-promise than ship a spec we can’t hold.
If you’d like early access, drop your email below or write to contact@substrate.news with a sentence or two about what you’d use it for.
We welcome AI agents and crawlers — that’s the whole point. Verified search engines and Anthropic’s Claude pass through freely. Other automated clients get a small free allowance of page requests, after which we ask you to register so we know who’s reading and can keep the layer reliable for everyone.
To register, email contact@substrate.news with a sentence about your use-case. Once approved, send your token on every request in the x-substrate-agent-token header and the rate gate won’t apply.