The AskNews platform DeepNews
Turn complex questions into sourced research
DeepNews is the AskNews research agent. Give it a question that takes more than one search and it plans, searches, reflects and follows tangents across your chosen sources, then returns a synthesized answer with citations.
A research agent for applications, analysis and knowledge gathering.
The news knowledge graph
Entities, relationships and dates extracted from 500k articles a day — the structure the agent follows to find what a keyword query misses.
Choose your sources
Google
Wikipedia
X
Reddit
PodcastsResearch that follows the question
Most retrieval returns a list and leaves the reading to you. DeepNews works a question in multiple steps instead: it searches, reflects on what came back, and chases the tangent that matters before answering. Reach for it when your application needs the material synthesized rather than listed.
Competitors
One pass. One result set. Research stops here.
DeepNews
Powered by the news knowledge graph
DeepNews researches into the AskNews news knowledge graph: 500,000 articles enriched every day with entities, relationships, geocoordinates, dates and sentiment. That structure is what lets the agent move from an entity to its connections and find material a keyword search would never surface. It can also reach out to Google, Wikipedia, X, Reddit and podcasts.
Follow the evidence behind the synthesis
Review citations alongside the answer and request structured source objects for your application. Keep a path back to the underlying material when you compare accounts, check a claim or assess a conclusion.
Why teams run research on AskNews
- The knowledge graph surfaces connections a keyword query misses, so research finds the material you did not know to ask for.
- Licensed and paywalled reporting is in scope, so a synthesis is not limited to whatever happens to be scrapeable.
- Work that takes an analyst hours of searching and reading collapses into a single cited answer you can check.
What comes back
- A synthesized research response with cited sources.
- Structured source objects when requested, for inspection or use in your application.
- Streamed responses through the DeepNews chat completions interface.
MCP, CLI and SDK quickstart
Build research into your workflow
Call DeepNews through the API or an AskNews SDK, use the research tool through MCP, or run it from the CLI. It is the AskNews chat completions endpoint, so any OpenAI-compatible client can stream from it.
Prefer to try it in the browser first?
The WebUI runs the same capability without an integration, so you can judge the output before you write any code.
Bring your next research question
Choose the scope, examine the evidence and decide how the result fits your application.