Skip to main content
Research runs inside brand chat and returns findings to the same thread.

​
Research lanes

  • External public research for companies, markets, competitors, websites, trends, public profiles, and public social posts.
  • Connected account insights for first-party performance questions across linked social accounts.
  • TikTok discovery research for niche, creator, trend, and content exploration when the research service is available.
  • Public video links for the metadata, captions, and recent uploads of public videos, channels, and playlists when the video metadata service is available.
Use /research for non-SEO research. Use /seo for page audits, URL comparisons, schema, content gaps, and AI search readiness. Research tasks often appear as specialist rows in chat while daoco gathers sources or connected-account data. Click a row to follow the research in its own view. Results stream back into the same thread when the run completes. If a requested report has no specialist result, daoco makes one more attempt. A completed or provisional report counts as a result; a report still waiting for approval stays paused. If no result arrives after the retry, the conversation marks the report as missing.

​
Search public social conversations

Name one or more platforms when you want daoco to find public social posts, videos, or discussions about a topic.
daoco checks an eligible connected API first. When that route is unavailable or has no matching results, the social research preview can use a bounded server-side broker or platform-constrained public-web search. Broker coverage currently focuses on X, YouTube, Reddit, Instagram, LinkedIn, and TikTok; actual availability can vary by workspace and platform. A public social search may take up to two minutes while those sources and the web fallback finish. A failed search does not establish that there were no mentions; review the coverage status alongside the findings. The inline research surface shows:
  • Live search progress and elapsed time while sources are checked.
  • A row for each requested platform with its source category and result count.
  • Clear No results, Unavailable, or Failed states for platforms without usable coverage.
  • Successful results from available platforms even when another requested platform fails.
Broker-backed searches use public content only. daoco does not pass your connected social-account credentials to the broker, and each broker request is bounded.

​
How research picks a data source

At the start of each research turn, daoco checks what this brand can reach: its connected social accounts, Search Console and Analytics, other connected tools, a linked LinkedIn browser session, and how many browser reads the run has left. Research then uses the most direct source for each question. Questions about your own accounts and site use connected data. Research never swaps in public estimates for your own numbers; if a connection is missing, it says what to connect. When you give research a link, it picks the reader for that kind of link. Social posts go to the social post reader, LinkedIn profiles to the LinkedIn reader when the brand has a LinkedIn browser session linked (without one, to public profile data sources and search results), and other pages to a fast page fetch. If a source is blocked or comes back empty, research names the next options, such as another source or a connection you could add, instead of retrying the same one.

​
Pages that need a real browser

Research reads a webpage with a fast page fetch first. When that fetch fails, is blocked, or returns almost no text, which is common for sites built with JavaScript, research opens the page once in a real browser and reads its main content, with no extra step from you. Each browser read can take up to two minutes, and a research run makes at most three. The browser opens signed out and only reads. It does not click, sign in, or fill in forms. If the page asks for a login or shows a bot check, research reports that the page isn’t publicly readable and moves on rather than retrying. LinkedIn profiles are the exception. If the brand has a LinkedIn browser session linked in Connections, research reads profiles and LinkedIn posts through it while signed in. Without one, LinkedIn shows research a sign-in wall, so research goes straight to public profile data sources and search results, and tells you that linking a LinkedIn browser session would give it signed-in reads.

​
Research structured public data

When daoco’s dedicated research tools do not cover a structured public-data request, research can check Monid’s catalog for a suitable source. The research agent may do one focused web-search pass first, but it will not keep rewriting the same search when the missing answer is a field or record that webpages do not expose. Common uses include public professional and company records, firmographic data, product catalogs, listings, and reviews. daoco inspects the selected source before running it, verifies that it is read-only, and limits the request to a small result set. It skips Monid when a connected tool or an authoritative webpage already answers the question. Chat looks up a single person or company record itself, without starting a research step. For example, it can read the LinkedIn profile of someone you want to email before the writer drafts the message. SEO work uses the same records for a competitor’s company details or an author’s profile. Larger lookups, such as many records or a comparison, still go to research. Monid lookups are paid from the workspace balance, so one request gets at most five of them. The specialists it hands work to share those five. Once they are used, daoco answers from the records it already has and searches public webpages for anything still missing. Browsing the catalog and checking a source’s inputs are free and don’t count. Completed Monid steps show the returned source links directly beneath the tool call. Use those links to review the evidence behind public profiles, company records, and other structured findings.
Monid is for public research. It does not receive connected-account credentials and cannot publish, message, purchase, schedule, or change external data. Private, residential, leaked, and personal-only contact data are outside the research agent’s permitted scope.

​
Read public video links

Paste a public video, channel, or playlist link and daoco can read its public metadata and captions: title, description, creator, follower count when shown, upload date, duration, view, like, and comment counts, tags, chapters, and a transcript from the captions. For a channel, profile, or playlist it lists recent videos with the counts the platform shows.
Supported links are public YouTube, TikTok, Instagram, X, Facebook, Reddit, Vimeo, Twitch, and Pinterest content, plus LinkedIn video posts. daoco cannot read LinkedIn profiles, Threads, or private, age-restricted, or login-only content. Counts are public snapshots at the time of the read, not your connected-account analytics.
daoco reads metadata and captions only. It does not download or store the video, and it does not use your connected-account credentials. Platforms sometimes block signed-out reads; when that happens daoco says the link could not be read instead of guessing.

​
Start research

Research uses basic tokens. Longer or multi-step research may consume more than a simple chat response.
Connected-account research depends on provider permissions and available analytics. LinkedIn personal profiles may return account-level insights without importing native posts, while LinkedIn pages and other supported accounts can use richer import and post-analysis flows when available.