Lanseq Company Classification & Extraction avatar

Lanseq Company Classification & Extraction

Pricing

from $1.68 / 1,000 extracted texts

Go to Apify Store
Lanseq Company Classification & Extraction

Lanseq Company Classification & Extraction

Classify company descriptions for lead segmentation and extract fields from supplied text. Use your own categories; export JSON/CSV with source evidence, missing values and review flags. Includes a free n8n workflow. No separate model API key required.

Pricing

from $1.68 / 1,000 extracted texts

Rating

0.0

(0)

Developer

Lanseq

Lanseq

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

3 days ago

Last modified

Categories

Features

Limited permissions

Share

Classify company descriptions with your own labels and extract fields directly supported by the supplied text. Get structured results with evidence, missing-value flags and JSON/CSV export.

Lanseq is a multi-open-model inference provider.

What you can do

  • Categorize company descriptions by industry or business activity.
  • Extract source-backed fields from text you already have.
  • Identify records that need review instead of filling missing information with guesses.

No separate model API key is required.

Quick start

Paste an input like this, set a maximum charge for the run, and select Start:

{
"items": [
{"id": "company-1", "text": "ExampleCo manufactures industrial bearings."}
],
"labels": ["manufacturing", "software", "logistics"],
"fields": ["companyName"]
}

Open Output when the run finishes and export JSON or CSV. Use only labels you want returned. Field values are verbatim text spans or null; this Actor does not normalize addresses or infer undisclosed facts.

Pricing

TaskFree / StarterScaleBusiness
Classification only, per 1,000 texts$0.98$0.98$0.98
Field extraction, with optional classification, per 1,000 texts$1.88$1.78$1.68

Each successfully stored and structurally validated result triggers one event. Extraction already includes any requested classification, so the two prices are never added together. Platform usage is included. There is no startup fee. Technical failure rows are not charged.

A valid result can contain null values or a review flag and is still billable: the price is for processing and delivering the result, not a guarantee that every requested fact exists. Set a maximum charge per run; the Actor stops before processing an item it cannot charge within that budget.

Input limits

  • 1–20 records per run, each with a unique id and up to 8,000 UTF-8 bytes of text.
  • Up to 20 labels and 10 fields; provide at least one label or field.
  • Apify Free-plan accounts can process up to 10 model requests per account in total, subject to shared daily availability. This allowance does not reset by creating a new run.
  • Paid-plan accounts use available capacity at the listed prices. Busy capacity produces a technical failure without a result fee.
  • Model requests are sequential. No automatic model retry occurs.

The free-account allowance is an ongoing cost-control policy, not a time-limited product launch.

Reading the output

Each row includes the record id, label, source evidence, extracted fields, review flag, usage, latency and a request id. Missing or unsupported values are null. Review flagged records before relying on them.

Structure and source-span validation do not establish semantic correctness. The evidence can confirm what the supplied text says, not whether the underlying company claim is true. This service does not verify purchase intent, contact validity, certifications or financial facts against external databases.

Technical errors stop the run, preserving completed rows. Inspect the Dataset and run summary before submitting unfinished records in a new run. Resurrecting a run with existing output is refused to avoid duplicate processing.

Data handling

Only submit text you are permitted to process. Input text is sent to Lanseq for inference. Apify stores run input and output under its platform settings. The gateway records request identifiers, pseudonymous account identifiers, timing and token usage; it does not log customer prompts or model response content.

Full input text is omitted from Dataset output unless includeSourceText is enabled. Evidence and extracted fields may still contain portions of that text. No zero-retention or independently audited accuracy guarantee is made.

Support

Contact jc@lanseq.cloud with the run ID and a description of the issue. Do not send passwords, API keys or confidential source material.