clay-cpj

name	clay-cpj
description	Use this skill when the user wants to search Clay companies, people, or jobs via Clay CPJ CLI, build lead/account/job lists, run Clay source searches, export results, or inspect Clay tables.
user-invocable	true

Clay CPJ CLI Skill

Clay CPJ CLI is a local CLI wrapper around Clay's Companies, People, Jobs source APIs.

It supports:

clay-cpj companies search ...
clay-cpj people search ...
clay-cpj jobs search ...

Credentials and sessions are stored locally in ~/.clay-cpj/credentials.json and ~/.clay-cpj/session.json. Set CLAY_WORKSPACE_ID explicitly for workspace-specific work.

Search Commands

Companies

clay-cpj companies search \
  --countries "United States" \
  --industries "Software Development" \
  --limit 25 \
  --output json

People

clay-cpj people search \
  --domains "openai.com,anthropic.com" \
  --title-keywords "Engineer,Developer Advocate" \
  --mode full \
  --cleanup \
  --output csv \
  -f /tmp/people.csv

Jobs

clay-cpj jobs search \
  --domains openai.com \
  --title-keywords Engineer \
  --max-num-days-since-posted 30 \
  --output json

Shared Flags

All entity searches support:

Flag	Notes
--mode preview/full/auto	Preview is fast and capped; full creates a Clay table
--limit N	Client/source result limit
--output csv/json/sqlite	Entity-shaped outputs
--output-file, -f	Output path
--cleanup	Delete generated Clay table/workbook after extraction
--wait-timeout N	Seconds to wait for full-mode Clay runs before returning pending
--detach	Create the Clay table and exit immediately without waiting
--quiet, -q	Suppress progress
--inputs-json	Exact raw Clay source inputs
--inputs-file	JSON file with exact raw Clay source inputs

Raw inputs replace typed filter construction. Only mode, output, cleanup, and client-side limit remain CLI-controlled.

Full-mode searches persist workbook, table, view, and source IDs to ~/.clay-cpj/runs/ as soon as the Clay table is created. A wait timeout means the remote Clay run is still pending, not failed.

Company Table Joins

People and jobs can start from an existing Clay company table by extracting domains:

clay-cpj people search --from-company-table t_xxx --company-domain-field Domain --title-keywords CEO
clay-cpj jobs search --from-company-table t_xxx --company-domain-field Domain --title-keywords Engineer

Outputs include source_company_table_id and source_company_domain when this join path is used.

Table Commands

clay-cpj table list
clay-cpj table info <table_id>
clay-cpj table count <table_id>
clay-cpj table export <table_id> --entity jobs --output sqlite -f /tmp/jobs.db
clay-cpj table delete <table_id>

Rules

• Use --mode full --cleanup for production searches that must be complete.
• Use --detach or a short --wait-timeout for long-running full-mode searches, then export with table export after Clay finishes.
• Use preview for filter validation and small smoke tests.
• Use --inputs-json or --inputs-file when exact Clay UI payload parity matters.
• Do not commit credentials, cookies, session files, workspace-specific artifacts, or exported Clay data.