Sign in once. Siftree reads your records and reacts to new events as they happen.
List Documents
Lists documents for a given extractor by reading extraction records (GET /v1/extracted_data/{extractorId}) and returning their documentId and fileName , deduplicated by documentId . AlgoDocs has no dedicated documents…
List Extractors
Lists all extractors in the authenticated AlgoDocs account (GET /v1/extractors). Each extractor includes at least id and name . Use this to discover a valid extractorId before running Upload File or List Documents .
List Folders
Lists all folders in the authenticated AlgoDocs account (GET /v1/folders). Each folder includes at least id and name . Use this to discover a valid folderId before running Upload File .
New Extracted DataTrigger
Emit new event for each newly extracted data record for a given AlgoDocs document (polls GET /v1/extracted_data/{documentId}). Each extraction record's stable id is used for deduplication. Run the List Documents action…
New Table Row ExtractedTrigger
Emit new event for each individual table row extracted from an AlgoDocs document (polls GET /v1/extracted_data/{documentId}). AlgoDocs represents a record's extracted data as a flat object, with any table/repeating…
Act
Take action in AlgoDocs
Agents update records and send messages for you, automatically, as drafts, or after you approve.
Upload File
Uploads a local file to an AlgoDocs folder for processing by a specific extractor via multipart/form-data (POST /v1/document/upload_local/{extractorId}/{folderId}). Returns the created document record including its id…
Includes its id and upload metadata. Run List Extractors to find a valid extractor ID and List Folders to find a valid folder ID before calling this action. The returned document id is the value to configure in New Extracted Data, and more
Siftree is an AI platform that sales, CX, marketing, product, and strategy teams use to find, win, and keep customers.
Siftree connects calls, tickets, reviews, social, documents, video, and the systems you already use — then builds pipelines, datasets, tools, and cited insights your team and agents can operate on. Long-horizon agents keep building after you walk away.
- Describe what you need in plain language; Siftree builds pipelines, structures the data, and routes models for language, vision, audio, transcription, classification, and enrichment.
- Operate on datasets, tools, reports, and cited insights. You define the rules — what counts, how it's scored, what gets labeled.
- Marketplace: 14,824 sources across 4,169 apps and websites as of October 2026, plus any public website. Search, choose a source, and create an analysis-ready dataset without scraper maintenance.
- Native connectors for help desks, CRMs, and call recorders, plus upload and API. Audio and video are transcribed and labeled with everything else.
- Runs on a schedule or a trigger, and alerts your team in Slack when a threshold is crossed.
- Acts in your systems: writes labels and fields back to the help desk or CRM, opens issues in Linear or Jira, and drafts replies and emails. Each agent runs automatically, drafts for review, or asks first, and every action is logged.
- Ask questions of your data from Claude or ChatGPT through Siftree MCP.
- Enterprise protections: continuous verification of every request, encryption in storage and transit, privilege and identity governance, immutable activity logs.
- Teams use the eight products in place of separate point tools for prospecting data, support analytics, social listening, community triage, ad research, AI search visibility, research repositories, and competitive intelligence.