@fluid-app/fluid-cli-connect

Fluid CLI plugin for Fluid Connect: link a back-office platform (Exigo, InfoTrax, ByDesign, or Pillars) to Fluid, then set its credentials, data mappings, and sync settings.

npm install -g @fluid-app/fluid-cli @fluid-app/fluid-cli-connect
fluid login
fluid connect list                                   # each provider and whether it's installed
fluid connect install exigo --yes                    # install the droplet, creating the connection
fluid connect credentials exigo --yes \
  --set api.company=acme --set api.username=api_user --set api.password=... \
  --set mssql.dataserver=sql.example.com --set mssql.database=acme \
  --set mssql.username=reader --set mssql.password=...
fluid connect show exigo                             # state, supported mappings, every setting
fluid connect unmapped exigo --auto-map --yes        # map exact name matches, list what's left
fluid connect mappings exigo customer_type --options # Fluid's and Exigo's values to map
fluid connect map exigo customer_type --yes --set source_type_slug=rep --set target_type_id=2
fluid connect settings exigo --yes --defaults        # take every default
fluid connect settings exigo --yes --set order.refund_handling=hold
fluid connect status exigo --range 24h               # sync health: success rate, tier, failures by category

Set up interactively

fluid connect walk exigo     # or: fluid setup run connect-exigo

walk runs the whole setup in the terminal, one question at a time:

1. Installs the provider if it isn't installed, after you confirm.
2. Asks each credential (passwords and tokens are typed hidden; Exigo also asks the environment, and the sandbox ID outside production). Fluid tests the connection before saving; if the provider refuses, its reason is shown and the questions are asked again, with Enter keeping what you typed. On a connected provider, you can keep the saved credentials.
3. Maps what's unmapped. Exigo first offers to map exact name matches, then gives a searchable list of Exigo's values; for the others you type the provider's value. Answer * to send every remaining value of that type to one value. Ranks and ship methods get a note and an optional row by hand.
4. Settings: keep them as they are, put them all back to the defaults, or go through the common settings, each with its label, description, default, and options.
5. Prints a summary, and the result as JSON.

Every answer is saved as you give it, through the same calls as the commands above. Type < to go back (a changed mapping replaces the row it saved), > to skip, and ? for the menu: jump to a part, change an answer, finish, or stop. Run walk again later to pick up what's left. In a script, use the commands above with --yes; walk needs a terminal.

Commands

unmapped lists, per mapping type, the Fluid values that have no row and no * fallback. done: true means nothing is left. Ranks and ship methods come back as check_by_hand because the API doesn't list Fluid's values for them. --auto-map works for Exigo only, since it is the one provider whose values the API lists. It adds a row wherever a Fluid value's name matches exactly one Exigo value, and leaves ambiguous names for a person.

A new connection already starts on the default settings. settings --defaults puts them back, all at once or one at a time with --only order.refund_handling.

Values passed with --set end up in your shell history and are visible in process listings. For passwords and tokens, prefer fluid connect walk, which asks for them without echoing. credentials keeps every value as typed (a password like 123456 stays a string), and creates the connection if installing the droplet hasn't yet. unmapped exits non-zero when a mapping type couldn't be checked.

status reads the provider's sync status from Fluid (/api/company/v2026-10/connect/integrations): the dashboard's success rate, tier, synced, held, quarantined and skipped counts, and the top failure categories, for --range 60m, 24h (the default), 7d or 30d. It carries counts and category codes, never error text. A provider the company hasn't installed prints the same result as list. When Fluid can't read the status, it exits non-zero with the error's status and code: 404 not_found, 409 not_connected (fluid-integrations has no record of the installation), 502 backend_unavailable or backend_rejected, or 503 connect_unconfigured (Fluid has no API host for the integration). It needs the droplets.view permission.

Every command prints JSON; failures print a JSON error on stderr and exit non-zero. A failed credential test prints the provider's reason in details. Commands that change live configuration ask for confirmation unless you pass --yes.

The CLI talks to two services with the same Fluid API token: api.fluid.app to install the droplet and read status, and fluid-integrations for everything else. Set FLUID_INTEGRATIONS_BASE to point at another fluid-integrations host.

The walkthroughs for each provider, including data mapping and the common settings, are the Connect guided setups at https://docs.fluid.app/setup.