Skip to content

Research Assistant

Use Deep mode Internal agents with Web search and Knowledge to find and read sources and produce structured reports.

Typical use: Competitive analysis brief → agent searches the web, reads sources, cites findings, and delivers a markdown report.


Research is time-consuming and tedious:

  • Information overload - Too many sources, not enough time
  • Manual synthesis - Hours spent reading and summarizing
  • Inconsistent quality - Different researchers, different approaches
  • Slow turnaround - Research takes days or weeks
  • Limited coverage - Can’t monitor everything relevant

The cost of slow research:

  • Missed opportunities from delayed insights
  • Poor decisions based on incomplete information
  • Researcher burnout from repetitive work
  • Competitive disadvantage
  • Inability to scale research operations

Research agents find and read sources and write up what they found, on demand or on a schedule.

Multi-Source Gathering

Search web, academic papers, news, social media, internal docs, and databases simultaneously.

Intelligent Synthesis

Analyze findings, identify patterns, extract key insights, and generate summaries.

Scheduled monitoring

Track topics on a schedule with a Loop, and get a digest when something new appears.

Citation Management

Automatically cite sources, track references, and generate bibliographies.

Competitive Intelligence

Monitor competitors, track industry trends, and identify market opportunities.

Custom Reports

Generate formatted research reports with executive summaries and visualizations.


  1. Define Research Topics

    Specify what to research: keywords, questions, sources to monitor, and depth of analysis needed.

  2. Agent Gathers Information

    Searches multiple sources, filters relevant content, and extracts key information automatically.

  3. Analysis & Synthesis

    Agent analyzes findings, identifies patterns, compares sources, and generates insights.

  4. Deliver Reports

    Receive formatted reports via email, Slack, or dashboard. Schedule daily, weekly, or on-demand.


Brief: “Summarize competitor pricing pages for products X, Y, Z.” The agent searches the web and reads each pricing page, then returns a structured markdown comparison with its sources.


The agent automatically:

  • Web search - Built-in web search over the open web
  • Publicly accessible papers - Preprints and open-access articles (gated databases like JSTOR/IEEE require your own access)
  • News monitoring - Public news pages and RSS feeds
  • Social & forums - Publicly accessible discussions
  • Internal sources - Knowledge attachments (Confluence, Notion, Google Drive, uploaded docs)
  • Competitor tracking - Websites, product updates, announcements

Beyond simple summarization:

  • Trend identification - Spot emerging patterns and themes
  • Sentiment analysis - Gauge public opinion and reactions
  • Comparative analysis - Compare sources and perspectives
  • Gap analysis - Identify missing information
  • Credibility assessment - Evaluate source reliability
  • Contradiction detection - Flag conflicting information

Professional, formatted reports:

  • Executive summary - Key findings and recommendations
  • Methodology - Sources and approach used
  • Detailed findings - Organized by theme or question
  • Data visualizations - Charts, graphs, timelines
  • Citations - Properly formatted references
  • Appendices - Raw data and supporting materials

Web search

Built-in web search covers the open web, and the agent reads the pages it finds, including many publicly indexed papers and preprints.

News & feeds

Monitor public news pages and RSS feeds, or wire a news API through Custom MCP with your own key.

Social & forums

Read publicly accessible discussions; gated platform APIs (X, Reddit, LinkedIn) require your own credentials via Custom MCP.

Web extraction

Extract data from websites and forums with the browser and scripts in the sandbox.

Internal sources (Knowledge)

Attach Confluence, Notion, Google Drive, a web crawler, or uploaded documents as Knowledge.

Reference export

Output citations as formatted text/BibTeX for import into your reference manager. No native Zotero/Mendeley connector.


Understand your market:

  • Customer needs and pain points
  • Market size and growth trends
  • Competitive landscape analysis
  • Pricing and positioning research

Stay ahead of competitors:

  • Product launches and updates
  • Marketing campaigns and messaging
  • Hiring patterns and expansion
  • Customer reviews and sentiment

Accelerate scholarly work:

  • Literature reviews
  • Citation discovery
  • Methodology research
  • Data collection and analysis

Due diligence and analysis:

  • Company research and financials
  • Industry trends and forecasts
  • Risk assessment
  • Regulatory changes

Build better products:

  • User needs and feedback
  • Feature requests and complaints
  • Technology trends
  • Best practices and patterns

Research automation scales with your needs:

  • Free: $0/month — 1,000 credits
  • Pro: $29/month — 15,000 credits ($24/mo billed annually)
  • Business: $99/month — 60,000 pooled credits ($83/mo billed annually)
  • Enterprise: Custom pricing for high-volume research

See Pricing for full plan limits and credit rates. Credit usage depends on research depth and the actions the agent takes.


  1. Sign up free - No credit card required

  2. Follow the quickstart - Create your first agent in 10 minutes

  3. Use the web-research Skill - attach it to your Agent in the Console

  4. Add internal sources - Confluence, Notion, Google Drive, or files as Knowledge



  • How accurate is the research? Agents cite all sources and flag conflicting information. You control which sources to trust.
  • Can it access paywalled content? If you have institutional access, agents can use your credentials to access paywalled databases.
  • What about data privacy? Research runs and their results stay in your organization unless you have the agent send them somewhere.
  • Can it handle specialized topics? Yes. Provide domain-specific sources and terminology, and agents adapt to any field.
  • How current is the information? Web search reads the live web when the agent runs. To follow a topic, run the agent on a schedule with a Loop.

View all FAQs →