@archiveye/cli

Command-line interface for the ArchivEye Developer Platform.

Install

npm install -g @archiveye/cli

Configure

archiveye config set --api-key aek-your-key-here
# production MCP host (not the BFF origin):
archiveye config set --api-key aek-your-key-here --mcp-base-url https://mcp.archiveye.ai
archiveye config show
archiveye whoami

Commands

| Command | Description | | --- | --- | | config set / config show | Save or display CLI configuration | | whoami | Authenticated employee profile | | collections list / create / update / delete | Collection lifecycle | | collections add-documents / remove-document / replace-document | Collection membership | | search -q "…" --collection-id ID | Semantic search via MCP | | documents upload / delete / status | Document lifecycle | | ingest start / status / cancel | Ingest jobs | | webhooks endpoints list / create / test | Webhook endpoints | | webhooks deliveries list / retry | Delivery log |

Retrieval and organization knowledge

archiveye search -q "ACME" --collection-id 12 --fulltext --document-ids 42,43
archiveye prepare-search -q "ricavi" --collection-id 12 --mode both
archiveye documents content 42
archiveye documents grep --document-ids 42,43 --pattern "ricavi" --fixed-string
archiveye documents read-range 42 --start-line 1 --end-line 40
archiveye tabular schema 42 --collection-id 12
archiveye tabular query 42 --collection-id 12 --sql "SELECT * FROM t LIMIT 10"
archiveye organization profile
archiveye organization directory --limit 100 --offset 0
archiveye knowledge search -q "budget" --limit 20
archiveye knowledge explore --limit 100
archiveye knowledge neighborhood NODE_ID --depth 2
archiveye knowledge status

knowledge search uses MCP to search stored company graph entities and relations with case-insensitive literal matching. It requires an employee key, Enterprise and organization workspace read access, but no collection ID or active Business OS. The MCP client can use the results to answer questions even if Business OS was disabled, removed or is in error. The separate knowledge ask -q "…" command uses the organization AI via REST and requires that AI configuration to be active.

New document sources

archiveye documents upload-directory ./reports --folder-id 10
archiveye documents upload-web --file article.md --source-url https://example.com/article
archiveye documents ingest-partner --file notes.md --title "Meeting notes" --source mnemo --external-id mnemo:note:42

Directory uploads create a fresh hierarchy, skip symlinks and empty subdirectories, and stop on the first error, leaving completed uploads in place. They require an employee key. Partner ingest updates the same document when its external ID is reused. Use --help on any command for all options.

Documentation

Full reference: https://data.archiveye.ai/docs/cli

License

MIT