deepsieve-cli

Run cited deep research from your terminal.

uv tool install deepsieve-cli   # or: pipx install deepsieve-cli
deepsieve login

login opens your browser, you approve a short code, and the CLI stores a scoped API key. Nothing to copy, nothing pasted into your shell history. The credential appears in Settings → API keys on your deployment and can be revoked there at any time.

The key covers the CLI's own commands and cannot create a Blueprint. If an agent will use it to create Blueprints, run deepsieve login --preset agent; the approval screen lists the extra scopes.

Commands

deepsieve whoami                              # identity, organization, active Blueprint, scopes
deepsieve runs create --query "..." --dry-run # free simulated run (~15s)
deepsieve runs create --blueprint <id> --query "..." --wait  # real run: spends credits
deepsieve runs create --blueprint <id> --seeds "..." --allowed-source "*.gov"  # only cite these
deepsieve runs create --blueprint <id> --query "..." --expected-total 1000  # refuse (charge nothing) if it now costs more than the estimate
deepsieve runs list
deepsieve runs get <id>                       # exit 4 while still running
deepsieve runs cancel <id>
deepsieve members list                        # who is in the organization + seat usage
deepsieve changes list --live                 # merges, renames, successor links
deepsieve changes undo <id>                   # asks first; --yes in scripts
deepsieve merge preview <entity> --survivor <id> --loser <id>   # writes nothing
deepsieve merge apply <entity> --survivor <id> --loser <id> [--choice KEY=loser] --preview-digest <digest>
deepsieve runs retry <id>                     # a retry that charges shows its total and asks
deepsieve runs reruns <id>                    # a report's re-research waiting for approval
deepsieve runs monitoring <id> --off          # stop a report's monitoring; asks first
deepsieve links remove <entity> <record-id> <linked-entity> <target-id>   # asks first
deepsieve subscription cancel                 # at period end; nothing new is charged; asks first
deepsieve subscription resume                 # undo it while the period runs; asks first
deepsieve data catalog                        # entities + columns (never guess)
deepsieve data get companies --receipts       # rows with per-cell citations
deepsieve data history companies headquarters_city --record <id>   # one cell's change history
deepsieve chat thread <run-id>                # changes, with status + warnings
deepsieve chat confirm <action-id>            # apply a change that waits for confirmation
deepsieve chat dismiss <action-id>            # decline it; nothing is written
deepsieve chat revert <action-id>             # undo a change a chat answer applied
deepsieve setup mcp                           # MCP registration for this origin

Add --json to anything for machine-readable output.

Profiles: several deployments side by side

deepsieve login --profile staging --origin https://staging.deepsieve.ai
deepsieve --profile staging runs list

Each profile is its own (origin, credential) pair, so staging and production never overwrite one another. DEEPSIEVE_PROFILE sets the default.

CI and containers

Skip login entirely — set DEEPSIEVE_API_KEY (and DEEPSIEVE_BASE_URL if not production). Environment variables take precedence over any stored profile.

DEEPSIEVE_API_KEY=ds_live_... deepsieve --json data get companies

What it deliberately cannot do

No billing, no organization or member management, no API-key management, and it cannot edit an active Blueprint. Those are decisions for a human in the app — and the restriction is enforced server-side by the credential's scopes, not just by the absence of a subcommand.

Starting a real (billable) run asks for confirmation, and refuses outright in a non-interactive shell unless you pass --yes.

Exit codes

Full docs: https://deepsieve.ai/developers/cli