Server facts
talonic-mcp
Project
- Author
- talonicdev
- Repository
- github.com/talonicdev/talonic-mcp
- License
- Not specified
- Category
- Files & Documents
Tech
- Runtime / language
- Not specified
- Transports
- stdio
- Remote available
- No
- Authentication
- Not specified
Activity
- GitHub stars
- 5
- Last commit
- Data as of
Tags
- Structured Data
- anthropic
- document extraction
- forms
- ocr
- schema
- talonic
Overview
Document workflows in Talonic center on extracting named fields rather than returning only OCR text. Agents can supply an inline schema or select a saved workspace schema, then submit file bytes, an accessible local path, a document URL, or an existing document identifier. Extraction returns JSON organized by that schema, confidence information for individual fields, and document details such as detected type and language. Optional provenance supplies source text, section, and page references for extracted fields. Schema listing and saving support repeated extraction across documents with the same structure.
Workspace retrieval separates conceptual discovery from structured matching. Search finds relevant documents, fields, sources, and schemas, while filtering selects documents using conditions on previously extracted values. Field types matter: numeric comparisons depend on numeric schema definitions. A metadata tool provides details about an individual document, including its processing log. For tasks requiring document text instead of selected fields, the Markdown tool retrieves OCR output for summarization, translation, or analysis. This gives agents distinct paths for locating documents, inspecting their contents, and extracting structured records.
talonic-mcp setup notes
- talonic-mcp records stdio only; confirm a network transport upstream before planning remote access.
- talonic-mcp has no recorded license; confirm reuse terms in its repository.