Plugin tracker

Tools

Explore what tracked Claude Connectors and ChatGPT Plugins can actually do. Search by tool, Plugin, Brand, category, verb, or access requirement.

Latest snapshot2026-09-12USmethodology registry-public-v1

Searchable tools

113,018

Authless tools

7,424

Auth required

100,766

Described tools

61,167

113,018 tools

Get Upload URLgetuploadurl · Get a pre-signed URL to upload a file to the LlamaParse S3 storage. Not available when authenticating with an API key — use uploadFileByUrl instead.LlamaParseLlamaIndexPluginrequiredAI
Grep File in Indexgrepfilefromindex · Grep the content of a file from an index, providing its file ID, the pattern to grep for and, optionally, a number of context characters and a maximum number of grep matches to retrieveLlamaParseLlamaIndexPluginrequiredAI
List Directorieslistdirectories · List the directories in a project. Returns one page at a time — pass the returned nextPageToken to fetch more. By default only user directories are listed; indexes create their own internal output directories, which are not valid sources for a new index.LlamaParseLlamaIndexPluginrequiredAI
List Directory Contentslistdirectory · List the files inside a directory, along with the directory itself. Returns one page at a time — pass the returned nextPageToken to fetch more. Use directoryFileId, not fileId, when referring to a file in later calls.LlamaParseLlamaIndexPluginrequiredAI
List Indexeslistindexes · List all the available indexes on the LlamaParse Platform. Indexes are vector-indexed directories with the possibility of searching/reading/grepping files and performing retrieval.LlamaParseLlamaIndexPluginrequiredAI
List Projectsgetuserprojects · List the projects available to the user, with their names, so you can pass the right projectId to other tools. Use this whenever a tool needs a projectId and the correct project is not already known — pick by name, and ask the user if the name is ambiguous.LlamaParseLlamaIndexPluginrequiredAI
Parse Fileparsefile · Parse a file providing its file ID, retrieving markdown or plain text content of the file. Use with file IDs obtained with the getUploadUrl/uploadFileByUrl tool or that the user providedLlamaParseLlamaIndexPluginrequiredAI
Parse File with LiteParseparsewithliteparse · Parse a PDF file with LiteParse, a fast, in-process parser that does not consume credits from the LlamaParse Platform. The tool needs a file ID obtained with the getUploadUrl/uploadFileByUrl tool or provided by the user. Only works with PDF files.LlamaParseLlamaIndexPluginrequiredAI
Read File from Indexreadfilefromindex · Read the content of a file from an index, providing its file ID and, optionally, an offset and a maximum length (in characters) to read.LlamaParseLlamaIndexPluginrequiredAI
Search Indexretrievefromindex · Perform hybrid search on the index, providing a query and, optionally, the top K documents to retrieve and the top N documents to rerank. Results carry provenance (file ID, page range) and, when the index stores them, references to per-page screenshots.LlamaParseLlamaIndexPluginrequiredAI
Search Schema Templatessearchschematemplates · Search the built-in library of starter extraction schemas (invoice, contract, resume, 10-K, patient intake and more) by keyword or category. Returns matching templates with their top-level field names, but not the full JSON Schema — call `getSchemaTemplate` for that. Prefer this over `generateExtractionConfig` when the user wants a common document type: it is instant and needs no sample file.LlamaParseLlamaIndexPluginrequiredAI
Split Filesplitfile · Split a file into category-based segments providing its file ID. Use with file IDs obtained with the getUploadUrl/uploadFileByUrl tool or that the user providedLlamaParseLlamaIndexPluginrequiredAI
Sync Indexsyncindex · Re-index a directory, picking up files added or changed since the last run. Indexes do not refresh on their own, so this is the only way an existing index sees new documents. Runs in the background — poll getIndexStatus until status is ready. If a sync is already running, this returns syncStarted=false (the underlying API responds 409 or 429): do not retry syncIndex, call getIndexStatus to check the running sync and wait until status is ready.LlamaParseLlamaIndexPluginrequiredAI
Upload File by URLuploadfilebyurl · Upload a file to LLamaParse S3 storage providing a URL to download the file data. On upload completion, the file will be sent to LlamaParse S3 storage, so that it can be used for downstream processing tasks like parsing, classification or splitting.LlamaParseLlamaIndexPluginrequiredAI
Add Sheet Tabadd · Add a new tab to an existing Google Sheet. Use this when you need an additional sheet after creation. Returns the new tab's sheetId.ShareWatchShareWatchPluginrequiredSecurity
Block File Accessblock · Block MCP access to a file. Use when a user says they shared a file by mistake or wants an agent's access stopped. Blocking gates access; it does not un-share the file, so the name and ID can still appear in listings. There is no unblock tool — the user restores access from the ShareWatch dashboard, so tell them that before blocking.ShareWatchShareWatchPluginrequiredSecurity
Build Presentationbuild · Create a Google Slides deck from a structured outline. Unlike create_presentation, which takes raw slides, this one applies Google's own layouts, so it suits decks meant to look designed. Pick a layout per slide and fill only the slots that layout provides; Google positions everything via placeholders (no manual geometry). Layouts are Google's own PredefinedLayout names, so they match the Slides API you already know. Layout → slots: TITLE = title + subtitle (opening slide); SECTION_HEADER = title only (a divider — no subtitle); TITLE_AND_BODY = title + bullets OR body; TITLE_AND_TWO_COLUMNS = title + body/bullets (LEFT) + body_right/bullets_right (RIGHT) — fill both sides or the empty one shows as a blank box when edited; MAIN_POINT / BIG_NUMBER / TITLE_ONLY = title only; BLANK = nothing. Set bullets for lists (it wins over body when both are given); keep each slide to one idea. Workflow: build, then call describe_slides to catch dropped content / overflow / overlap, then preview_slides for a final visual check.ShareWatchShareWatchPluginrequiredSecurity
Connection Statusget · Report this user's ShareWatch setup status: the authenticated user's email, workspace connection status, default folder, org membership, and available actions. Setup problems surface here as a clear status, whereas they surface in other tools as unrelated-looking failures. When multiple ShareWatch connectors are configured, this identifies which Google account this connector is authenticated as.ShareWatchShareWatchPluginrequiredSecurity
Convert to Google Formatconvert · Convert an Office or PDF file to the matching Google-native format: .pptx → Google Slides, .xlsx/.csv → Google Sheets, .docx/.pdf → Google Docs. Creates a new file — the original is unchanged. ALWAYS ask the user before converting.ShareWatchShareWatchPluginrequiredSecurity
Create Doccreate · Create a new Google Doc in the user's Drive. The content is inserted as plain text — Markdown is NOT rendered, so headings and bold written here arrive as literal ** and # characters. To produce a formatted document, create it and then call format_document, which converts Markdown into real Google Docs styling.ShareWatchShareWatchPluginrequiredSecurity
Create Foldercreate · Create a folder in Google Drive. Use this to organize files. The folder can be set as the default location for new files.ShareWatchShareWatchPluginrequiredSecurity
Create Presentationcreate · Create a new Google Slides presentation with optional slides.ShareWatchShareWatchPluginrequiredSecurity
Create Spreadsheetcreate · Create a new Google Sheet. Pass sheet_names_json to create multiple tabs at once (data_json writes to the first tab).ShareWatchShareWatchPluginrequiredSecurity
Delete Filedelete · Permanently delete a file from Google Drive. ALWAYS confirm with the user before deleting — this cannot be undone.ShareWatchShareWatchPluginrequiredSecurity
Format Doc from Markdownformat · Write markdown content into a Google Doc with full formatting — headings, bold, italic, bullets, numbered lists, links, and horizontal rules. Use this instead of create_document or update_document when the content has structure that should be visually formatted. Pass file_id to replace an existing doc's content, or omit to create a new doc.ShareWatchShareWatchPluginrequiredSecurity
Inspect Slide Layoutdescribe · Inspect the layout of every slide without rendering images. Returns a geometry-level readback — each element's position, size and text, plus flags for content that overflows the slide bounds, overlaps another element, or uses a sub-legible font. Cheap (no image render) — use it to catch mechanical layout problems while editing; use preview_slides for the final visual pass.ShareWatchShareWatchPluginrequiredSecurity
List Fileslist · List the Google Workspace files ShareWatch can see. This is not the user's Drive — it is only what was created through ShareWatch plus what the user has explicitly shared with it, so a short list means a narrow grant and an empty one does not mean an empty Drive. Never tell the user a file does not exist on this basis; a file you cannot see here may still be there. Returns file IDs, names, types, URLs, and each file's parent folder — folders appear as entries too, so parent_folder_id is what tells you which files live inside which folder.ShareWatchShareWatchPluginrequiredSecurity
Move Filemove · Move a file into a different folder. Both the file and destination folder must be accessible to ShareWatch.ShareWatchShareWatchPluginrequiredSecurity
Open File by Linkaccess · When a user shares a Google Drive URL or asks you to access a specific file, use this tool. Accepts any Google Drive, Docs, Sheets or Slides URL and resolves it against the Google Drive API (https://developers.google.com/drive/api/reference/rest/v3). If the file is already accessible, returns its content. If not, returns a grant link the user can open to select it. If the user mentions the file's name, pass it as file_name_hint so the picker can search for it.ShareWatchShareWatchPluginrequiredSecurity
Preview Slidespreview · Get thumbnail images for all slides in a Google Slides presentation. Returns rendered PNG images of each slide. Use this for a final visual check of layout and formatting.ShareWatchShareWatchPluginrequiredSecurity
Read Fileread · Read the text content of a Google Doc, Sheet, or Slides presentation. Docs export as plain text, Sheets as CSV, Slides as plain text. For large files, returns a size warning instead of content — use max_chars and offset to read in chunks. file_id also accepts a ShareWatch handoff code of the form SW-XXXXXXXX. The user gets one by opening a file from Google Drive with ShareWatch, and it stands for that file for one hour. The result always reports the real file_id, so use that value for any follow-up call rather than the code.ShareWatchShareWatchPluginrequiredSecurity
Report a Bugreport · Tell the ShareWatch developers what this server got wrong. Use it for outright bugs — an unexpected error, a wrong result — and equally for friction: a confusing parameter, an unhelpful error message, a task that took three calls when it should have taken one. Your assessment of the API is the point; it is the only way its designers learn how it behaves in practice. The report is written to the user's own audit log, visible to them and to their workspace administrators, and to ShareWatch's operational logs. Write about the tool, never about the user's documents or conversation.ShareWatchShareWatchPluginrequiredSecurity
Request File Accessrequest · When a user wants you to access a file that isn't in the file list, call this tool. It returns a URL the user can open to select and grant access to files from their Google Drive.ShareWatchShareWatchPluginrequiredSecurity
Search Filessearch · Search for files by name, text content, type, or date range. Use this instead of list_files when the user has many files or is looking for something specific. Use content when the user remembers what a file says but not what it is called. All parameters are optional — combine them to narrow results.ShareWatchShareWatchPluginrequiredSecurity
Set File Read-Onlyset · Mark a file as read-only. Read-only files can be read but not updated through ShareWatch. This is enforced at the proxy layer regardless of Google Drive permissions. One direction only: read_only=false is refused, because only the user can make a file writable again, from the ShareWatch Files page.ShareWatchShareWatchPluginrequiredSecurity
Update Docupdate · Update an existing Google Doc. Supports replacing all content or appending to it. The result reports the mode that was actually applied and how many characters were written, so you can confirm a replace was a replace.ShareWatchShareWatchPluginrequiredSecurity
Update Presentationupdate · Modify an existing Google Slides presentation. Structured operations (operations_json): add_slide, insert_text, delete_text, delete_slide, replace_text, update_text_style, update_page_bg. Google's own spellings are accepted where the operation is identical — replace_all_text and create_slide both work. An unrecognised type, or one missing a required field, is REJECTED with a message naming the operation; nothing is silently skipped. Examples: - delete_text: {"type":"delete_text","element_id":"id"} - replace_text: {"type":"replace_text","old_text":"foo","new_text":"bar"} - update_text_style: {"type":"update_text_style","element_id":"id","bold":true,"font_size":24,"color":"#1a56db"} - update_page_bg: {"type":"update_page_bg","slide_id":"id","bg_color":"#1a1a2e"} Raw mode (raw_requests): Pass a Google Slides API BatchUpdatePresentationRequest body directly for full API access. Request shapes are documented at https://developers.google.com/slides/api/reference/rest/v1/presentations/batchUpdateShareWatchShareWatchPluginrequiredSecurity
Update Spreadsheetupdate · Write data to a range in an existing Google Sheet. The range must reference a tab that already exists (e.g. 'Overview!A1' requires an 'Overview' tab). Errors with 'Unable to parse range' if the tab name is unknown — call add_sheet_tab first to create new tabs.ShareWatchShareWatchPluginrequiredSecurity
Upload Fileupload · Upload a binary file (xlsx, pdf, docx, png, etc.) to Drive from base64-encoded content. Use this when you have file bytes from the conversation (e.g. an attached spreadsheet) and need to put them in the user's Drive. Set convert=true to have Drive convert Office formats to native Google formats during upload. Maximum 20MB decoded.ShareWatchShareWatchPluginrequiredSecurity
List Documentslist · Start here. Lists every knowledge-base document this user may read - webinar write-ups, newsletter issues, procedures and guides - as a flat array with `parentId`, so the tree can be reconstructed. Returns titles and slugs only; call read_document for a body. Some entries have `type: "FOLDER"` - they only group other entries and have no content of their own; do not try to summarise one, read its children instead. Each entry carries a short `description` - use those, plus the tree shape, to decide what is worth reading, then call search_documents for a question or read_document for a slug you already have. Never answer from a name or description alone. The list is already scoped to the user's access, so anything absent from it is not readable and should not be guessed at.Analytics BrainWitBeePluginrequiredData & Analytics
Read Documentread · Returns one document in full, as markdown, together with its ancestor chain for context. Takes the `slug` from list_documents or search_documents. If `type` comes back as "FOLDER", there is no body - it only groups other entries; use list_documents and filter by `parentId` to see what it contains. Cite the document by name when you use its content, so the reader can find the source. A slug the user has no access to comes back as an error, not as empty content - do not retry it.Analytics BrainWitBeePluginrequiredData & Analytics
Search Documentssearch · Full-text search across the titles AND bodies of the knowledge-base documents this user may read. Returns matches with a snippet around the hit; call read_document for the full text. Prefer this over list_documents when the question is about what the documents say rather than what exists. Matching is a plain substring match, so search for distinctive words rather than whole questions - two or three narrow searches beat one long one.Analytics BrainWitBeePluginrequiredData & Analytics
add_prospects_to_campaignadd · Queue LinkedIn profile URLs into the user's prospecting machine. Each profile is fetched, evaluated against the campaign's ideal customer profile (filter_mode 'icp'; 'all' skips the filter; 'custom' uses custom_filter_instructions), deduped against existing leads, and good prospects become new leads that MimikFlow then engages automatically (invite, personalized DM, follow-ups, AI conversation) at its normal LinkedIn-safe pace. Runs in the background. Accepts URLs from search_prospects or any other source (CRM export, spreadsheet, webinar attendee list). When the user names a SPECIFIC person to add ('add John Doe to the pipeline'), use filter_mode 'all': the ICP filter is meant for bulk sourcing and could reject an individually chosen contact. A 1st-degree profile with an ONGOING LinkedIn conversation is skipped (reported in the import summary): to hand a live thread to the AI, use list_manual_conversations + adopt_conversation instead. Confirm the list with the user before queuing. IMPORTANT for agencies running several clients: pass campaign_id to land the prospects in the RIGHT client's campaign (they are also evaluated against THAT campaign's ICP). Omitting it defaults to the primary campaign, so always set it when the user is working on a specific client -- call list_campaigns first if unsure which id.MimikFlowMimikFlowPluginrequiredSales & CRM
adopt_conversationadopt · Hand a manually-started LinkedIn conversation over to MimikFlow (the machine automating the user's LinkedIn prospecting): creates the contact as a lead in the campaign, imports the full message history, and from then on the machine works the thread under the campaign's settings. handled_by 'setter' (default): the AI answers replies, sends spaced reminders on silence and pushes toward the conversion goal. handled_by 'me': the thread is only tracked (visible in the inbox, replies surfaced), no AI message ever. CAUTION with 'setter': if the contact's last message is still unanswered the AI will answer it on its next run, and reminders resume on silence -- say this to the user and get an explicit confirmation per conversation before calling. Get chat_id from list_manual_conversations.MimikFlowMimikFlowPluginrequiredSales & CRM
approve_pending_messageapprove · Approve a pending draft: it is sent to the prospect on LinkedIn through the normal pipeline (daily LinkedIn caps apply) and the lead's status advances accordingly. Read the conversation first with get_conversation.MimikFlowMimikFlowPluginrequiredSales & CRM
drop_leadsdrop · Take prospects out of the active pipeline by setting their status to 'dropped': MimikFlow (the machine automating the user's LinkedIn prospecting) stops ALL automated outreach to them -- no more reminders, no AI replies -- while the lead, its conversation history and the LinkedIn connection are kept. This is the right way to 'remove' e.g. competitors or bad fits from the pipeline (hard deletion only exists in the app). Safety: if a dropped prospect writes back later, the reply is still surfaced to the user. DESTRUCTIVE-ISH: always list the exact leads (names + ids) to the user and get an explicit confirmation BEFORE calling. Any pending drafts for these leads are cancelled. Max 50 ids per call (from list_leads).MimikFlowMimikFlowPluginrequiredSales & CRM
edit_and_approve_pending_messageedit · Replace a pending draft's text with an improved version, then send it. Keep the user's voice; write in the same language as the conversation.MimikFlowMimikFlowPluginrequiredSales & CRM
export_leadsexport · Export a slice of leads with full fields (for pushing into a CRM, a spreadsheet, or reporting). Same filters as list_leads, higher limit. Returns the MOST RECENTLY updated leads only: when `has_more` is true, `total` matching leads exist and the user should narrow the slice with `status` or export from the app. Occasional use (a few exports a day); recurring extraction is not what this connector is for.MimikFlowMimikFlowPluginrequiredSales & CRM
get_account_overviewget · Global state of the user's MimikFlow account (MimikFlow = a machine that automates their LinkedIn B2B prospecting 24/7: finds prospects, sends invites, personalized DMs, follow-ups, and an AI setter that answers prospects to book meetings). Returns campaigns, LinkedIn session health, which pipelines are on/off, lead counts per status, inbox items waiting for the user (drafts to approve + escalated conversations), and the LinkedIn safety limits in force. Reading the pipeline states: the extra lead sources (engagement_scraper = reactions to the user's posts, profile_views = profile visitors, invite_manager = incoming invites, first_degree_scanner = their existing 1st-degree network) are OPT-IN and start off for everyone -- an off source is usually a setup choice to suggest enabling, EXCEPT the ones listed in `locked_sources`: those are unavailable on the user's current plan (e.g. first_degree_scanner needs a PAID Pro+ plan, it is not included in the free trial) -- never suggest enabling a locked source, present it as what unlocks when they upgrade/subscribe. tracking always runs. The pipeline ids in the payload are internal: to the user, call each step by its product name in their language (la recherche de prospects / prospect search, les demandes de connexion / connection requests, le premier message / the first message, les relances / follow-ups, l'IA qui répond aux prospects...), never by its id. Call this first to orient yourself.MimikFlowMimikFlowPluginrequiredSales & CRM
get_campaign_profileget · Read a campaign's current settings, i.e. everything that drives how the AI finds prospects and writes to them: targeting (my_target = ideal customer in plain text, my_anti_target = who to avoid), who_am_i (who the user is), my_product (their offer), the conversion goal (goal_type 'call' = book a meeting, 'link' = drive to a URL), the booking `calendar` block (connected / provider / applies_from / auto_book), voice (communication_style, style_examples, social_proof_stories, my_objections = the user's answers to common objections), and OPTIONAL behavior overrides (custom_cold_dm_instructions / custom_follow_up_instructions / custom_setter_instructions, setter_custom_stop_condition). IMPORTANT on booking: when `calendar.connected` is true, an EMPTY goal_link is perfectly normal and NOT a misconfiguration -- the AI reads the user's real availability and books the slot itself, so no link is needed. Only report a missing booking link when the goal is 'call' AND no calendar is connected. IMPORTANT: an empty custom field is normal, NOT a gap: MimikFlow's default behavior (varied reminders, respecting a no, stopping on silence, escalating to the human) applies on its own; these fields only personalize it further. The key names are internal: to the user, speak of each field by its plain meaning in their language ('ta cible', 'ton offre', 'tes réponses aux objections'), never by its raw key. ALWAYS read this before updating anything.MimikFlowMimikFlowPluginrequiredSales & CRM

Tool Explorer FAQ

What is Tool Explorer?

Tool Explorer indexes the callable tool names and descriptions attached to public registry profiles. It is useful for seeing what agents can actually invoke, not just which profile exists.

How do category and verb filters work?

Category filters use the live registry category rollup. Verb filters use the public tool insights rollup, so the page stays backed by the same read models as the tracker charts.

Why do auth requirements matter?

Auth requirements show whether a tool is likely usable without account connection, requires authentication, is private, or is unknown in the current snapshot.