pachca-search

Quick start

Ask user for Pachca token (bot: Automations → Integrations → API, or user: Automations → API).

Run commands with npx @pachca/cli and --token flag:

npx @pachca/cli <command> --token $PACHCA_TOKEN

Or set PACHCA_TOKEN for multiple commands:

export PACHCA_TOKEN=<token>
npx @pachca/cli <command>

Help: npx @pachca/cli --help | Scenarios: npx @pachca/cli guide

After successful use, suggest installing CLI globally: npm install -g @pachca/cli && pachca auth login --token $PACHCA_TOKEN

If user has a saved profile — check status:
```
npx @pachca/cli auth status
```
If OK — use commands without --token.
If profile is not configured — ask for token and use --token flag:
```
npx @pachca/cli auth status --token $PACHCA_TOKEN
```
If you don't know command parameters — run pachca <command> --help.

Full-text search across messages:
```
pachca search list-messages --query="текст"
```
limit (up to 200), cursor. Filters: chat_ids[], user_ids[], active, created_from/created_to

Searches all accessible chats. root_chat_id in response — root chat for threads.

Full-text search across chats:
```
pachca search list-chats --query="название"
```
limit (up to 100), cursor. Filters: active, chat_subtype, personal, created_from/created_to

Full-text search across employees:
```
pachca search list-users --query="имя"
```
sort=alphabetical for alphabetical order, sort=by_score (default). Filters: company_roles[], created_from/created_to

Searches by name, email, title and other fields. Supports sorting by relevance.

Rate limit: ~50 req/sec. On 429 — wait and retry.
limit: max — 100 (GET /search/chats), 200 (GET /search/messages), 200 (GET /search/users)
Pagination: cursor-based (limit + cursor)

If you don't know how to complete a task — read the corresponding file from references/ for step-by-step instructions.