Extend MCP

Create a workflow

create_workflow

Create a workflow — a multi-step document pipeline (workflows group): parse → extract/classify/split → validations → human review. name alone creates an empty draft; steps builds the graph up front (call get_documentation with https://docs.extend.ai/workflows/configuring-workflows.md before hand-authoring a step graph). The draft is the only mutable surface — edit with update_workflow, freeze with deploy_workflow_version, run with run_workflow.

Input Schema

TableJSON Schema

Name	Required	Description
`name`	Yes	Display name for the workflow (1-255 chars).
`steps`	No	Step graph (max 125 steps), TRIGGER → PARSE first. Every step needs { type, name }; "name" is REQUIRED and is what other steps route to. Route via next, an ARRAY of objects: linear steps (TRIGGER/PARSE/EXTRACT) use next: [{ step: "<target name>" }]; CLASSIFY/SPLIT branch with next: [{ step, classificationId }] (classificationId = a classification id from the config, not its type). TRIGGER routes to exactly one PARSE. Types: TRIGGER, PARSE, EXTRACT, CLASSIFY, SPLIT, MERGE_EXTRACT, CONDITIONAL, CONDITIONAL_EXTRACT, EXTERNAL_DATA_VALIDATION, WEBHOOK_RESPONSE, RULE_VALIDATION, VALIDATION, ROUTER, HUMAN_REVIEW, COLLECT, FILE_CONVERSION. EXTRACT/CLASSIFY/SPLIT need a config with exactly one of a saved ref or inline config (EXTRACT: config.extractor {id,version} or config.extractorConfig with REQUIRED schema; CLASSIFY: config.classifier {id,version} or config.classifierConfig); next is only allowed once config is set. Classifier/splitter refs can't be "latest" — use semver or "draft". The rules here are a summary — before authoring a step graph by hand, call get_documentation with https://docs.extend.ai/workflows/configuring-workflows.md and follow it.
`environment`	Yes	"TEST" = the Test (development) environment, "PRODUCTION" = live. Must match a granted target from get_me (an API key pins one environment).
`workspaceId`	Yes	Target workspace (ws_...). Must be a granted workspace — get_me lists the accepted values.

Output Schema

TableJSON Schema

Name	Required	Description	Default
`id`	Yes
`name`	Yes
`createdAt`	No
`updatedAt`	No
`draftVersion`	No

Schema Changelog

Changes observed during successful MCP inspections.

Changed1 schema field changedOct 2, 2026

changedInput schema / properties / steps / description

Previous value: -"Step graph (max 500 steps), TRIGGER → PARSE first. Every step needs { type, name }; \"name\" is REQUIRED and is what other steps route to. Route via next, an ARRAY of objects: linear steps (TRIGGER/PARSE/EXTRACT) use next: [{ step: \"<target name>\" }]; CLASSIFY/SPLIT branch with next: [{ step, classificationId }] (classificationId = a classification id from the config, not its type). TRIGGER routes to exactly one PARSE. Types: TRIGGER, PARSE, EXTRACT, CLASSIFY, SPLIT, MERGE_EXTRACT, CONDITIONAL, CONDITIONAL_EXTRACT, EXTERNAL_DATA_VALIDATION, WEBHOOK_RESPONSE, RULE_VALIDATION, VALIDATION, ROUTER, HUMAN_REVIEW, COLLECT, FILE_CONVERSION. EXTRACT/CLASSIFY/SPLIT need a config with exactly one of a saved ref or inline config (EXTRACT: config.extractor {id,version} or config.extractorConfig with REQUIRED schema; CLASSIFY: config.classifier {id,version} or config.classifierConfig); next is only allowed once config is set. Classifier/splitter refs can't be \"latest\" — use semver or \"draft\". The rules here are a summary — before authoring a step graph by hand, call get_documentation with https://docs.extend.ai/workflows/configuring-workflows.md and follow it."New value: +"Step graph (max 125 steps), TRIGGER → PARSE first. Every step needs { type, name }; \"name\" is REQUIRED and is what other steps route to. Route via next, an ARRAY of objects: linear steps (TRIGGER/PARSE/EXTRACT) use next: [{ step: \"<target name>\" }]; CLASSIFY/SPLIT branch with next: [{ step, classificationId }] (classificationId = a classification id from the config, not its type). TRIGGER routes to exactly one PARSE. Types: TRIGGER, PARSE, EXTRACT, CLASSIFY, SPLIT, MERGE_EXTRACT, CONDITIONAL, CONDITIONAL_EXTRACT, EXTERNAL_DATA_VALIDATION, WEBHOOK_RESPONSE, RULE_VALIDATION, VALIDATION, ROUTER, HUMAN_REVIEW, COLLECT, FILE_CONVERSION. EXTRACT/CLASSIFY/SPLIT need a config with exactly one of a saved ref or inline config (EXTRACT: config.extractor {id,version} or config.extractorConfig with REQUIRED schema; CLASSIFY: config.classifier {id,version} or config.classifierConfig); next is only allowed once config is set. Classifier/splitter refs can't be \"latest\" — use semver or \"draft\". The rules here are a summary — before authoring a step graph by hand, call get_documentation with https://docs.extend.ai/workflows/configuring-workflows.md and follow it."

Changed1 schema field changedOct 1, 2026

changedInput schema / properties / steps / description

Previous value: -"Step graph (max 100 steps), TRIGGER → PARSE first. Every step needs { type, name }; \"name\" is REQUIRED and is what other steps route to. Route via next, an ARRAY of objects: linear steps (TRIGGER/PARSE/EXTRACT) use next: [{ step: \"<target name>\" }]; CLASSIFY/SPLIT branch with next: [{ step, classificationId }] (classificationId = a classification id from the config, not its type). TRIGGER routes to exactly one PARSE. Types: TRIGGER, PARSE, EXTRACT, CLASSIFY, SPLIT, MERGE_EXTRACT, CONDITIONAL, CONDITIONAL_EXTRACT, EXTERNAL_DATA_VALIDATION, WEBHOOK_RESPONSE, RULE_VALIDATION, VALIDATION, ROUTER, HUMAN_REVIEW, COLLECT, FILE_CONVERSION. EXTRACT/CLASSIFY/SPLIT need a config with exactly one of a saved ref or inline config (EXTRACT: config.extractor {id,version} or config.extractorConfig with REQUIRED schema; CLASSIFY: config.classifier {id,version} or config.classifierConfig); next is only allowed once config is set. Classifier/splitter refs can't be \"latest\" — use semver or \"draft\". The rules here are a summary — before authoring a step graph by hand, call get_documentation with https://docs.extend.ai/workflows/configuring-workflows.md and follow it."New value: +"Step graph (max 500 steps), TRIGGER → PARSE first. Every step needs { type, name }; \"name\" is REQUIRED and is what other steps route to. Route via next, an ARRAY of objects: linear steps (TRIGGER/PARSE/EXTRACT) use next: [{ step: \"<target name>\" }]; CLASSIFY/SPLIT branch with next: [{ step, classificationId }] (classificationId = a classification id from the config, not its type). TRIGGER routes to exactly one PARSE. Types: TRIGGER, PARSE, EXTRACT, CLASSIFY, SPLIT, MERGE_EXTRACT, CONDITIONAL, CONDITIONAL_EXTRACT, EXTERNAL_DATA_VALIDATION, WEBHOOK_RESPONSE, RULE_VALIDATION, VALIDATION, ROUTER, HUMAN_REVIEW, COLLECT, FILE_CONVERSION. EXTRACT/CLASSIFY/SPLIT need a config with exactly one of a saved ref or inline config (EXTRACT: config.extractor {id,version} or config.extractorConfig with REQUIRED schema; CLASSIFY: config.classifier {id,version} or config.classifierConfig); next is only allowed once config is set. Classifier/splitter refs can't be \"latest\" — use semver or \"draft\". The rules here are a summary — before authoring a step graph by hand, call get_documentation with https://docs.extend.ai/workflows/configuring-workflows.md and follow it."

First observedSep 3, 2026

TDQS

A4.6/5.0

Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint=false, idempotentHint=false, openWorldHint=true and destructiveHint=false. The description adds genuinely new behavioral context: the draft is the only mutable surface, and it must be frozen via deploy before running. It stops short of detailing permissions or error behavior, but the added routing/lifecycle model is real value.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Three dense sentences, front-loaded with the tool's identity and the two creation modes before the lifecycle and the documentation pointer. It is long relative to most definitions but nearly every clause carries a distinct instruction rather than restating the schema.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a tool that creates a complex graph-shaped resource with an output schema present, the description supplies the authoring model, the lifecycle hand-offs to sibling tools, and the authoritative doc link. Nothing an agent needs to invoke it correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline would be 3, but the description goes further: it explains the semantic weight of each parameter mode (name alone → empty draft vs. steps → full graph) and the ordering/authoring constraints for the step graph, which is the least obvious part of the schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb (Create) and resource (a workflow) and immediately defines what a workflow is as a multi-step pipeline with the stages enumerated. It also distinguishes itself from the mutate/deploy/run siblings (update_workflow, deploy_workflow_version, run_workflow) by contrast.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicitly says 'name alone creates an empty draft; steps builds the graph up front' and names the alternative (edit with update_workflow, freeze with deploy_workflow_version, run with run_workflow). It also gives a clear prerequisite for the harder path: call get_documentation before hand-authoring a step graph.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Extend MCP tools Inspect

Try in Browser

Shareable, User-Owned Context BackendPeople, applications, and agents read, write, and coordinate together in one context lake. Reach it over MCP, a CLI, a Python SDK, or REST.fulcradynamics.comAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?