One balance across your workspace
The web app, Chrome extension, CLI and AI connector share your research-credit balance. Ask Adlicio to check your credits, or use whoami, before planning a larger research session.
Connect Adlicio to your AI app. Find real customer language, explore ad angles, and use the research you already have.
MCP connects your AI app to your Adlicio account. No API key needed.
Choose where you work. We’ll show you the setup.
Open Claude’s connector settings.
https://mcp.tryadlicio.com/mcpUse this URL for manual setup.
Open Settings and enable Developer mode.
https://mcp.tryadlicio.com/mcpAuthentication: OAuth.
Your client needs HTTP and OAuth support.
https://mcp.tryadlicio.com/mcpThe same endpoint for every remote client.
Ask your AI app:
Check my Adlicio account and remaining credits.Uses whoami · No research credits
Your AI returns your Adlicio account, plan, and remaining credits. If the account is wrong or the tool is missing, use the troubleshooting steps.
Next steps
You don’t need to memorize tool names. Tell your AI what you want to learn and ask it to use Adlicio.
Research what people on Reddit find frustrating about meal planning. Show recurring problems with direct customer quotes and source links.Replace “meal planning” with your customer’s problem. Read the source quotes before turning patterns into hooks.
Connect / Claude
Connect once. Use Adlicio in Claude, Desktop, and Cowork.
Available on Free, Pro, Max, Team, and Enterprise. Free allows one custom connector. For Team and Enterprise, an Owner must add Adlicio first; each member then connects their own account.
Connect / ChatGPT
Add Adlicio as a custom app in ChatGPT on the web.
Requires Plus, Pro, Business, Enterprise, or Education with developer-mode access. Workspace permissions apply. Some workspaces use Settings > Apps > Advanced Settings and Apps > Create. Review action confirmations when prompted.
Connect / Cursor
Add a remote server to your Cursor configuration.
{
"mcpServers": {
"adlicio": {
"url": "https://mcp.tryadlicio.com/mcp"
}
}
}No API key or OAuth client secret is needed. Your organization may restrict which MCP servers you can connect. Cursor asks for tool approval according to your run settings.
Connect / Claude Code
Register the hosted server, then sign in from Claude Code.
claude mcp add --transport http adlicio https://mcp.tryadlicio.com/mcpThe command adds Adlicio for your current project. Add --scope user to make it available across projects. If you already connected through Claude, check /mcp before adding another configuration.
Connect / Other clients
Use a client that supports remote MCP with OAuth.
Setup labels and plan requirements vary by client. Adlicio supports dynamic OAuth client registration. Clients that only support local stdio servers need the separate CLI setup in the developer docs.
Reference
Your AI discovers these tools automatically. Browse by task, or use search to jump to a tool. Expand a row for its inputs and details.
Cost and access are different. A tool can use zero credits and still require a paid plan. Tool availability can also depend on the connected collection service.
research_topicSearch Reddit, collect matching discussions and rank customer language in one call. Use customer wording for the keyword and narrow the subreddits when the topic is broad. Add a brand to save the research to its workspace.
keywordbrandlimitsubredditsexclude_scrapedmodemax_commentsmin_scoremax_depthsortinclude_body_onlyscrape_urlCollect comments or reviews from one supported public page. Returns customer text with its source details. The price depends on the source; stored results can be reused free. It reads a URL rather than searching for one.
urlbrandforcemodemax_commentsmax_threadsmin_scoremax_depthsortinclude_body_onlysearch_redditFind relevant threads with their titles, URLs and snippets, without collecting comments. Threads already in your history include a scrape id that get_history_item can read for zero credits.
keywordlimitsubredditssearch_youtubeFind videos with their titles, channels, views and URLs. Use a keyword of 2 to 5 words, the way a viewer types it. Returns video listings, not comments. Follow with scrape_url to read a video's comments for zero credits, or start_deep_scrape for a larger collection.
keywordlimitordersearch_blindFind public Blind posts where professionals discuss a keyword. Returns titles, snippets, engagement and URLs, without collecting comments. Searches use the account's search allowance.
keywordlimitsortsearch_hackernewsFind Hacker News stories or the individual comments that match a keyword, with points, comment counts and item URLs. Follow with scrape_url on an item URL to read the whole thread for zero credits.
keywordlimitsorttypefind_scroll_stoppersFind public TikTok and Instagram videos for a niche or competitor, ranked by engagement, with their caption hooks. These are video and hook results; use research_topic or scrape_url for the customer comments behind them.
keywordplatformlimitfind_creatorsFind TikTok and Instagram creators with followers, engagement estimates, public contact details and the posts that surfaced them. Country is a ranked preference, not a guaranteed filter. Useful for UGC, affiliates and sourcing partners.
keywordplatformcountryminFollowersmaxFollowerslimitAvailability depends on the connected collection service. Reconnect your AI app to refresh its tool list.
Link to this toolfind_competitor_adsFind Facebook and Instagram ads in the public Meta Ad Library: copy, hooks, offers, CTAs and landing URLs. Filter or sort the fetched sample by known running duration. Longevity is a research signal, not proof of profit. Limited to 25 calls a day.
querycountryactive_statusmedia_typesort_bymin_days_runningcta_typecall_ads_onlykeyword_onlylimitAvailability depends on the connected collection service. Reconnect your AI app to refresh its tool list.
Link to this toolfind_tv_adsSearch the public iSpot creative catalog for ad copy, publication dates and video links. Catalog dates do not prove when an ad aired or whether it is running. Airing times, spend and audience targeting are unavailable. Limited to 25 calls a day.
querylimitAvailability depends on the connected collection service. Reconnect your AI app to refresh its tool list.
Link to this toolget_ad_transcriptGet the spoken transcript of one public Meta video ad using its numeric id or Ad Library URL. Transcriptions can contain errors; no on-screen text or performance is inferred. Image ads and unavailable audio cost zero credits. TV and arbitrary video URLs are unsupported.
adAvailability depends on the connected collection service. Reconnect your AI app to refresh its tool list.
Link to this toolget_video_transcriptGet the spoken transcript of one public TikTok video, Instagram reel or YouTube video or Short by URL, with timed cues where the platform gives them. Read what a creator or competitor says: the hook, the claims, the offer and the call to action. Pair it with scrape_url on the same link for the comments under the video. Costs 1 credit when a transcript is delivered. Videos with no speech, and Instagram videos over 120 seconds, cost zero credits.
urlAvailability depends on the connected collection service. Reconnect your AI app to refresh its tool list.
Link to this toolinspect_landing_pageRead a public landing page's HTML copy, headings, phone links, buttons, form labels and redirects. It does not submit forms or make calls. JavaScript-only content may be missing; blocked pages are not charged. Returns page evidence, not a screenshot or performance data.
urlfind_productsSearch Google Shopping for product titles, prices, merchants, ratings, review counts and URLs. Returns listings rather than review text. Use a supported product URL with scrape_url to read customer reviews.
querymax_resultsAvailability depends on the connected collection service. Reconnect your AI app to refresh its tool list.
Link to this toolfind_early_productsFind product listings showing early-adopter signals, with a score, supporting evidence and a signal breakdown. Search across marketplaces and customer discussions; use brand_filter and physical_only to narrow the results. An early signal is not proof a product will succeed.
nichemarketplacesprice_minprice_maxregionstrictnessbrand_filterphysical_onlymax_resultsAvailability depends on the connected collection service. Reconnect your AI app to refresh its tool list.
Link to this toolstart_deep_scrapeStart a background collection of available comments on one YouTube video, up to 20,000. Charged when a new job starts; an existing live job for the URL is reused without another charge. Check progress with get_deep_scrape.
urlmax_commentsforceget_deep_scrapeCheck a background YouTube collection using the job id from start_deep_scrape. When it finishes, read the stored comments with get_history_item using the returned scrape id. Wait at least a minute between progress checks.
job_idscrape_facebook_groupSearch a group you belong to through your own logged-in Chrome browser. Requires the Adlicio extension with “Let Claude use this browser” enabled. Returns a background job id; get_extension_job checks progress. Public Facebook posts use scrape_url instead.
group_urlkeywordmax_postsget_extension_jobCheck a queued extension job for a Facebook group or an Amazon pull handed to your browser. Reports browser connection status and the scrape id when complete. Read the saved comments with get_history_item; wait at least a minute between checks.
job_idwhoamiSee the connected Adlicio account, its plan and available research credits. A useful first connection check: it runs no research and uses no credits.
No inputs needed.
Link to this toollist_historyList saved scrapes and research from your dashboard, extension, CLI and AI chats, newest first. Filter by title, keyword or platform. Returns stored entries without collecting anything again.
queryplatformlimitoffsetget_history_itemRead one stored scrape or ad inspection by its history id, without fetching it again. Set full_text to true when you need untrimmed quotes or exports. Use offset and max_comments to read longer collections in pages.
idfull_textmax_commentsoffsetsearch_historyFind a word or short phrase across comments your account already collected. Each match includes its source scrape. Searches are case-insensitive; narrow the phrase or platform when the response says more matches are available.
queryplatformlimitget_brand_researchRead a brand's avatars, ad angles, customer quotes and saved scripts. With no arguments, lists the account's brands. Also reports the status of a run started with start_brand_research.
brandcreate_brandCreate a brand in the same workspace used by the Adlicio dashboard. Its audience, offer and positioning help shape later research. An exact name that already exists returns that brand instead of creating a duplicate.
nameurlnicheaudienceofferkey_claimtonemechanismstart_brand_researchStart the dashboard's full research pipeline for an existing brand: discussions, avatars, angles and supporting quotes. Charged when a new run starts. Usually takes 3–5 minutes; check its status with get_brand_research.
brandsave_scriptSave a finished script, hook set or advertorial into a brand's Scripts library. It appears in the dashboard alongside that brand's research. Use the brand's name or id; the title is optional.
brandscripttitlelist_tracked_sourcesList the listening sources already in your workspace, with their schedule, last collection and any reported error. Filter by brand to see one workspace's targets. Reading the list does not start a collection.
brandadd_tracked_sourceAdd a source to an existing brand for scheduled collection. Reddit targets are available on paid plans; other sources and keyword monitoring need All Access or Team. Target limits apply. Adding the target is free; later collections can use credits.
brandkindtargetcadenceget_audienceRead an audience's saved sources and recent research runs. With no arguments, lists your audiences. These are the communities and pages you grouped for research, rather than a generated list of customer personas.
audiencestart_voc_reportTurn 1–30 saved scrapes you own into a Voice-of-Customer report with themes, objections, angles and quoted evidence. Add product_context to tailor the report. Charged when a new report starts; use get_voc_report to read it when ready.
scrape_idsproduct_contextget_voc_reportRead a saved Voice-of-Customer report and its status using the report id. With no arguments, lists your recent reports. Works with reports created in the dashboard or through start_voc_report.
idReference / Sources
Give scrape_url a supported public URL. Reddit is available on every plan; other sources require All Access or Team.
A video URL, not a channel page.
Includes Stack Overflow questions.
Public pages only.
Product pages using Yotpo, Loox or Judge.me.
Stored or archived reviews. Have 2 credits available before starting an Amazon pull; only the delivered tier is charged.
Public threads and profiles. Free-plan preview limits apply.
Public posts. Private groups use the browser extension workflow.
Posts and reels. A pending result may need a later retry.
Product review pages, not individual /itm/ listings.
Charged only when the live collection tier delivers reviews.
A delivered review collection costs 2 credits.
Reference / Credits
Research credits pay for new collection and analysis. Check your balance in your account or ask your AI to use whoami.
The web app, Chrome extension, CLI and AI connector share your research-credit balance. Ask Adlicio to check your credits, or use whoami, before planning a larger research session.
A tool marked 0 credits does not spend your balance. Its plan label still applies: for example, YouTube search and history reads need All Access or Team even though they use no credits.
Free includes 6 research credits to start and a capped Reddit preview in chat. Those included credits do not reset monthly. All Access includes 250 credits a month; Team includes 750 shared credits a month. Purchased credits can affect the available allowance.
Metered searches and scrapes charge when a result is delivered. Brand research, Voice-of-Customer reports and deep YouTube collections charge when a new background job starts. Their progress checks use zero credits.
History reads use zero credits. Search results flag previously collected pages with a scrape id so get_history_item can read them without collecting again. Repeating the same URL scrape within 15 minutes does not add a second charge.
Have 2 credits available before starting an Amazon pull. Cache and archive results cost 0, a paid public teaser costs 1, and live reviews cost 2. The delivered result determines the charge.
All Access has daily fair-use limits of 200 Reddit searches and 150 URL scrapes. Meta ad searches and TV ad searches each have a 25-call daily limit. A remaining credit balance does not remove those limits.
Help
Start with the symptom you see. Your saved research stays in your Adlicio workspace when you reconnect.
Check that Adlicio is enabled in the current conversation. Then disconnect and reconnect it in your AI app’s connector settings to refresh the tool list. In a team workspace, ask the owner whether custom connectors are enabled.
Ask Adlicio to check your account with whoami. Disconnect the connector, then connect again and choose the same account you use in the Adlicio app.
Some collections run in the background. For a deep scrape, use get_deep_scrape with the returned job id. For Instagram, retry the same URL later. Brand research and Voice-of-Customer reports have their own progress tools.
Read the response to see whether the limit is your balance, daily allowance, or plan access. Check the tool’s plan label and credit cost. A zero-credit tool may still need All Access or Team.
Use the Adlicio Chrome extension in your own signed-in browser. The scrape_facebook_group tool queues the search; get_extension_job checks its status. Read the saved comments using get_history_item.
The hosted connector needs remote HTTP support. Use the commentscraper CLI’s local stdio MCP setup instead. It is a separate integration, described in the developer docs.
Technical details
https://mcp.tryadlicio.com/mcpFor local MCP, the CLI, REST API, and webhooks, read the developer documentation. You can disconnect Adlicio from your AI app at any time and revoke authorized devices in your account.