Skip to main content

talonic_list_schemas

List all saved schemas with their definitions. Returns schema IDs (both UUID and SCH-XXXXXXXX short format), names, and field definitions.

The talonic://schemas resource exposes the same data to clients that browse resources separately (Claude Desktop and Cowork render these in the UI).

`talonic_list_schemas` is a read-only tool that returns every saved schema in your workspace. Agents should call this before creating a new schema to avoid duplicates, and when the user asks about their available extraction templates. The response includes the full JSON Schema definition for each schema, so the agent can inspect field structures without additional calls.

Each schema in the response includes both a UUID (id) and a human-friendly short ID (short_id in SCH-XXXXXXXX format). Either can be passed as schema_id to talonic_extract. The short ID is easier for users to reference in conversation, while the UUID is useful for programmatic workflows.

Call talonic_list_schemas before talonic_save_schema to check if a similar schema already exists. This prevents cluttering the workspace with near-duplicate schemas.

Example

No input parameters required.

Tool response
{
  "schemas": [
    {
      "id": "sch_7a1b2c3d-4e5f-6a7b-8c9d-0e1f2a3b4c5d",
      "short_id": "SCH-A1B2C3D4",
      "name": "Standard Invoice",
      "description": "Extracts vendor, line items, totals, and payment terms.",
      "version": 3,
      "field_count": 6,
      "definition": {
        "type": "object",
        "properties": {
          "vendor_name": { "type": "string" },
          "invoice_number": { "type": "string" },
          "total_amount": { "type": "number" },
          "due_date": { "type": "string", "format": "date" },
          "currency": { "type": "string" },
          "line_items": { "type": "array" }
        },
        "required": ["vendor_name", "total_amount"]
      }
    },
    {
      "id": "sch_e5f6a7b8-c9d0-1e2f-3a4b-5c6d7e8f9a0b",
      "short_id": "SCH-E5F6A7B8",
      "name": "Certificate of Insurance",
      "description": "Extracts insured party, coverage types, limits, and expiry.",
      "version": 1,
      "field_count": 8,
      "definition": { "type": "object", "properties": { "..." : {} } }
    }
  ]
}

Example: check for duplicates before saving

Agent checks existing schemas before creating a new one
// User: "Create a schema for purchase orders"
// Agent first checks existing schemas:
// talonic_list_schemas → {}

// Response shows existing schemas:
{
  "schemas": [
    {
      "id": "sch_7a1b...",
      "short_id": "SCH-A1B2C3D4",
      "name": "Standard Invoice",
      "field_count": 6
    },
    {
      "id": "sch_c3d4...",
      "short_id": "SCH-C3D4E5F6",
      "name": "Purchase Order",
      "field_count": 7
    }
  ]
}

// Agent: "You already have a 'Purchase Order' schema (SCH-C3D4E5F6)
//  with 7 fields. Would you like to use it, or create a new one?"

Agents should call talonic_list_schemas proactively in several scenarios: before saving a new schema (to prevent duplicates), when the user asks about available extraction templates, and when starting a session that involves document extraction (to present schema options). The call is free and returns quickly, so there is no cost or latency penalty for checking frequently.

The version field on each schema tracks how many times it has been updated through the Talonic dashboard. A schema at version 1 has never been modified since creation. Higher versions indicate the schema has been refined — potentially adding new fields, changing types, or updating required fields. Agents can use this information to inform the user about schema maturity when choosing between multiple similar schemas.

The field_count in the response is a quick indicator of schema complexity without needing to parse the full definition. Schemas with fewer fields are faster to extract and produce more focused results. Schemas with many fields (10+) cover more ground but may have lower per-field confidence on complex documents. Agents can use field count to help users choose between a detailed schema and a simpler one for quick extractions.

Frequently asked questions

How do I list schemas via MCP?+
Call talonic_list_schemas to get all saved schemas with their definitions. The talonic://schemas resource also exposes this data.
What is the difference between the UUID and SCH-XXXXXXXX ID?+
Both reference the same schema. The UUID (id) is the full identifier, while SCH-XXXXXXXX (short_id) is a shorter, human-friendly format. Either can be passed as schema_id to talonic_extract.
Does talonic_list_schemas require any parameters?+
No. It takes no input parameters and returns all saved schemas in the workspace with their full definitions.
Is talonic_list_schemas free to call?+
Yes. It is a read-only metadata call that does not consume extraction credits. Agents can call it as often as needed to check for existing schemas, present options to the user, or verify a schema exists before using it in an extraction.
How do I know which schema to use for a document?+
Call talonic_list_schemas to see all available schemas with their names, descriptions, and field counts. Match the schema name and description to the document type. For example, use a 'Standard Invoice' schema for invoices and a 'Certificate of Insurance' schema for COIs. When unsure, start with an inline schema and save it after confirming the results.

Related

talonic_save_schema
talonic_extract