Classify company descriptions in batches with Apify Lanseq

Quick overview Classify existing company descriptions with your own labels. This workflow batches records through the Lanseq Apify Actor, preserves original IDs, and stops when results are incomplete, so missing records remain visible for review.

How it works Starts manually and loads the configured labels, budget caps, and input records. Validates record IDs and text size limits, enforces the execution opt-in and budget constraints, and splits the records into batches of up to 20 items. Submits the current batch to the Lanseq Actor on Apify as a single run with a per-run maximum charge cap. Polls the Apify run status every 10 seconds until it reaches a terminal state or a five-minute polling deadline is hit. When the run finishes, reads the Apify dataset items for that run and joins results back to the original input IDs, flagging missing, duplicate, or unexpected outputs. Continues with the next batch only if the previous batch succeeded and every row is structure-validated, then outputs a final report that includes all earlier results plus markers for any unsubmitted records.

Setup Select an existing Apify HTTP Header Auth credential in n8n on Start Actor once, Read run, and Read results. Configure authentication privately in n8n credentials; never put tokens in the workflow JSON or URL. In Input and budget, replace synthetic records with your own {id, text} items and set classification labels (maximum 20). Keep fields as an empty array for classification. IDs must be unique and at most 100 UTF-8 bytes; each description must be at most 8,000 UTF-8 bytes. Review the Actor pricing and build 0.1.3, plus the $0.05 per-run and $0.10 total cap defaults. The template is free; Actor execution may incur charges. Set executeModelCalls to true deliberately after reviewing the settings. Run manually and inspect Final report. Review missing, failed, duplicate or unexpected output and any unsubmitted records. On transport failure, inspect the saved n8n execution and Apify Runs before replaying: a second POST can create another charged run.

Requirements An n8n instance, an Apify account with access to the Lanseq Actor, an existing Apify authentication credential, and supplied company descriptions. No separate model-provider API key is needed. Source text is sent to Apify and the model service; INPUT and execution logs may retain it.

Customization Use your own business labels and map a stable ID plus a description from a CRM export or an upstream data workflow. This template does not crawl websites, discover contacts or enrich missing descriptions.

Additional info Validation: imported and executed in n8n 2.47.7 against an isolated localhost mock API. Checked complete 41-record processing and stopping after partial results; 12 adapter checks passed. The live Apify/model end-to-end path remains unverified. Structurally validated output is not a guarantee of classification accuracy. No pinned execution data or credentials are included. HTTP retries and Actor restart-on-error are disabled; a polling timeout does not cancel the remote run.

Setup and recovery guide: https://github.com/CHYYX/Bridge-ai/blob/lanseq-apify-n8n-20261115/examples/apify-company-classification/README.md

Actor: https://apify.com/everlasting_circlet/my-actor

0
Downloads
0
Views
7.48
Quality Score
beginner
Complexity
Author:Lanseq(View Original →)
Created:11/18/2026
Updated:11/18/2026

🔒 Please log in to import templates to n8n and favorite templates

Workflow Visualization

Loading...

Preparing workflow renderer

Comments (0)

Login to post comments