What can the Extend node actually do with a document?

It runs Extend's core jobs: parse a file into clean text, split a merged upload into separate documents, classify what each one is, and extract structured fields. You can trigger these one at a time or as batch runs, so a single workflow can take a raw upload all the way to routed, structured data.

Can a workflow react when an Extend run finishes?

Yes. The On Run Completed trigger starts a workflow the moment a classify, parse, split, extract, or workflow run finishes. That lets you fan the extracted output into other steps automatically instead of polling for the result yourself.

How do I connect Extend to NoClick?

Add your Extend API token in the node's credential panel and the connection is live. Every operation, from creating an extractor to launching an evaluation set run, uses that token, so you set it once and reuse it across the workflow.

Can it classify documents before extracting the fields?

Yes. Use the sync Classify operation or a classify run to identify the document type first, then branch the workflow so each type hits the right extractor. You can also create a classifier and publish new versions of it directly from the node.

Can I check accuracy before putting a change into production?

You can create an evaluation set, add items to it, and start an evaluation set run against a new extractor or classifier version. That gives you a measured comparison, so you promote a change only after the numbers hold up.

How does Codex get access to my apps?

You connect an integration to the agent's tools handle and select which operations to expose. Codex then calls those operations as tools, authenticated with the credentials you connect in NoClick.

Is this the same Codex CLI?

Yes. NoClick runs the Codex harness for you and adds your connected apps as tools. You do not have to install or manage anything locally.

Can I trigger a Codex agent automatically?

Yes. Because it runs hosted, you can start it from a schedule, a webhook, or another app event, not just by hand.

Which apps can I connect?

Any of NoClick’s tool-provider integrations, from Slack and Linear to GitHub, Notion, Google Sheets, Postgres, and dozens more.

Codex

Connect Codex to Extend

Give an AI agent your Extend account as a tool and it can read documents itself instead of handing them back to you. It can start an extract run on a file, parse and split a stack of pages, classify each document, and spin up a classifier or extractor on the fly, calling each operation as a tool while it works through a batch. It can even launch evaluation set runs to check accuracy before it acts on the results.

Loading preview…

Opens a workflow with 95 tools ready to use.

What happens next

Set up in minutes

The open button drops you into a guided setup. It asks exactly this, nothing else:

Choose where it runs
On your computer through your own Codex, or hosted in the cloud. The rest of setup adapts to your choice:
Pick what it can do in Extend
A preset of real Extend operations. Each one becomes a tool the agent can call, and it can't touch anything outside the list.
Connect Extend
One sign-in. The agent acts through your account, scoped to the operations you picked.
Add your personal MCP link
The last step mints a private MCP URL for this agent and walks you through exactly this, with your real link filled in. The page flips to Connected the moment your Codex first calls in.
Run this in your terminal
$codex mcp add noclick --url https://mcp.noclick.app/s/<your-link>
<your-link> is your personal server URL — setup mints it for you.

Exact capabilities

Tools Codex gets

Pick the operations you want and each becomes a tool the agent can call directly while it works.

Extend

95 tools

Upload File

Upload a document to Extend (multipart) and get a fileId for processing.

Get File

Retrieve a file's metadata, and optionally its parsed text/markdown/html.

List Files

List uploaded files (cursor pagination).

Delete File

Delete an uploaded file.

Parse (Sync)

Parse a document synchronously into clean markdown + blocks (≤5 min).

Extract (Sync)

Run structured extraction synchronously (≤5 min).

Classify (Sync)

Classify a document synchronously (≤5 min).

Split (Sync)

Split a multi-document file synchronously (≤5 min).

Edit (Sync)

Edit a document synchronously (≤5 min).

Create Parse Run

Parse a file into clean, chunked, LLM-ready markdown + structured blocks.

Get Parse Run

Poll a parse run for status and parsed output.

List Parse Runs

List parse runs (cursor pagination).

Cancel Parse Run

Cancel an in-progress parse run.

Delete Parse Run

Delete a parse run.

Batch Parse Runs

Create many parse runs in one batch.

Create Extract Run

Run structured data extraction on a file using an extractor.

Get Extract Run

Poll an extract run for status and extracted JSON.

List Extract Runs

List extract runs (cursor pagination).

Cancel Extract Run

Delete Extract Run

Batch Extract Runs

Create Extractor

Create a reusable extractor (schema + prompt config).

Get Extractor

Update Extractor

List Extractors

Create Extractor Version

List Extractor Versions

Get Extractor Version

Create Classify Run

Classify a file into one of a configured set of document types.

Get Classify Run

List Classify Runs

Cancel Classify Run

Delete Classify Run

Batch Classify Runs

Create Classifier

Get Classifier

Update Classifier

List Classifiers

Create Classifier Version

List Classifier Versions

Get Classifier Version

Create Split Run

Split a multi-document file into logical sub-documents.

Get Split Run

List Split Runs

Cancel Split Run

Delete Split Run

Batch Split Runs

Create Splitter

Get Splitter

Update Splitter

List Splitters

Create Splitter Version

List Splitter Versions

Get Splitter Version

Create Edit Run

Run a programmatic document edit (config = schema + instructions).

Get Edit Run

Delete Edit Run

Generate Edit Schema

Analyze a document and auto-generate an edit schema.

Get Edit Template

Create Workflow

Get Workflow

Update Workflow

List Workflows

Create Workflow Version

List Workflow Versions

Get Workflow Version

Create Workflow Run

Run a configured Extend workflow on a file.

Get Workflow Run

List Workflow Runs

Update Workflow Run

Cancel Workflow Run

Delete Workflow Run

Batch Workflow Runs

Create Webhook Endpoint

List Webhook Endpoints

Get Webhook Endpoint

Update Webhook Endpoint

Delete Webhook Endpoint

Create Webhook Subscription

List Webhook Subscriptions

Get Webhook Subscription

Update Webhook Subscription

Delete Webhook Subscription

Create Evaluation Set

List Evaluation Sets

Get Evaluation Set

Create Evaluation Set Items

List Evaluation Set Items

Get Evaluation Set Item

Update Evaluation Set Item

Delete Evaluation Set Item

Create Evaluation Set Run

Get Evaluation Set Run

Get Batch Run

Get Batch Processor Run

Starting points

What you can build

When a new PDF lands in your inbox or storage, start an Extend extract run and write the structured fields straight into your database.

When an Extend run completes, pick up the extracted data and route it to the right team based on the document type.

Classify every incoming file with a sync classify call, then branch the workflow by document type before extracting.

Split a merged scan of many documents into separate files with a batch split run, then classify and extract each one.

Batch parse a folder of contracts overnight and collect the parsed output for search and review.

Publish a new classifier version, run it against an evaluation set, and compare accuracy before you rely on it in production.

Meet the runner

About Codex

Codex is OpenAI's command-line coding agent. NoClick runs it as a hosted agent and lets you connect your apps to it as tools, turning it from a coding assistant into an operator that can act across your systems. Connect an integration to the agent and Codex can query, create, and update records in that app directly. You keep the Codex experience and give it real reach into the tools your team already uses.

Fast, capable code generation and editing

Good at well-scoped, concrete tasks

Strong at API and data manipulation

Familiar to teams already on OpenAI tooling

Good to know