Your data model is the structure of your CRM: your objects, their fields and choices, how they link, the tables and boards you saved, the layout of a record page, the dashboards you share and who may see what. You can save all of it as one file, keep it as a backup, move it to another workspace or hand it to a client as a ready-made template. The file holds the structure only: no records, no files and no secrets.
Export
- Open Settings, then Data model.
- Choose Export schema.
- Pick the objects to include, or leave everything ticked, and choose Download.
The file is called formspring-data-model.json. If you pick only some objects, the objects they are
linked to come along, so the file works on its own. Only the owner and administrators can export.
Import
- Open Settings, then Data model, and choose Import schema.
- Upload a file or paste its contents.
- Formspring compares it with your workspace and lists every change before anything happens: what would be created, updated, switched off or brought back, with the old and the new value of each part. Changes that need your confirmation are highlighted. Changes Formspring cannot make are marked and stop the import.
- Choose Apply.
Importing is all or nothing. If any part is refused (a plan limit, a view that names a field that does not exist), nothing is changed and you see why. Importing the same file twice changes nothing the second time.
By default, what your workspace has and the file leaves out stays. Tick Also switch off what the file leaves out to match the file exactly. Fields, relations and objects that are switched off keep their data and come back when you import a file that has them again; a saved view that the file leaves out is removed. A change that needs confirmation is one that:
- switches off a field that holds values, or an object that has records,
- removes choices from a select field (records that hold them keep them, but they are no longer offered),
- makes a field unique (it fails if two records already share a value).
A large file is applied in the background. The import screen shows its progress and what it did.
A field's type and the two ends of a relation never change. If the file has a different type for a field you already have, the import stops and names the field; give the new field another key and switch the old one off.
What the file looks like
A file is plain JSON. Everything is named by key (the key of an object, of a field, of a choice, the name of a view, the key of a role), never by an internal number, so the same file fits any workspace.
{
"format": "formspring.crm.schema",
"version": 1,
"objects": [
{
"key": "projects",
"kind": "custom",
"label_singular": "Project",
"label_plural": "Projects",
"label_field": "name",
"duplicate_criteria": ["code"],
"fields": [
{ "key": "budget", "label": "Budget", "type": "currency", "required": false, "unique": false,
"settings": { "default_code": "USD" } },
{ "key": "status", "label": "Status", "type": "select", "required": false, "unique": false,
"options": [
{ "value": "planned", "label": "Planned", "color": "gray", "default": true },
{ "value": "active", "label": "Active", "color": "green" }
] },
{ "key": "code", "label": "Code", "type": "text", "required": false, "unique": true }
],
"views": [
{ "name": "All Projects", "type": "table", "is_default": true },
{ "name": "Board", "type": "kanban", "group_by": "status",
"fields": [{ "field": "name", "visible": true }, { "field": "budget", "visible": true }],
"groups": [{ "value": "planned", "visible": true }, { "value": "active", "visible": true }] },
{ "name": "Big active projects", "type": "table",
"filter": { "operator": "and", "children": [
{ "field": "status", "operator": "is", "value": "active" },
{ "field": "client.region", "operator": "is", "value": "North" }
] },
"sorts": [{ "field": "budget", "direction": "desc" }] }
],
"permissions": {
"roles": { "editor": { "can_delete": false } },
"fields": { "budget": { "editor": { "can_read": true, "can_update": false } } }
}
},
{ "key": "clients", "kind": "custom", "label_singular": "Client", "label_plural": "Clients",
"fields": [{ "key": "region", "label": "Region", "type": "text", "required": false, "unique": false }] },
{ "key": "contact", "kind": "standard",
"fields": [{ "key": "tier", "label": "Tier", "type": "text", "required": false, "unique": false }] }
],
"relations": [
{ "kind": "one_to_many", "on_delete": "set_null",
"source": { "object": "clients", "key": "projects", "label": "Projects" },
"target": { "object": "projects", "key": "client", "label": "Client" } }
]
}
What each part holds:
| Part | What it says |
|---|---|
format, version |
The file type and its version. A file with a newer version than your workspace understands is refused. |
objects[].key, kind |
custom objects are created and updated. standard objects (contact, account, deal, task) are only referenced: their built-in fields are never described, but your own fields, views, permissions and layout on them travel. |
objects[].fields |
Your own fields in order: key, label, type, description, icon, required, unique, settings, default_value and, for choices, options with value, label, color and default. The built-in fields (name, owner, dates) are not listed. |
objects[].views |
The saved views everyone can open, of every type, with their columns, filter, sorts, grouping and board columns. A filter can use a field of a linked record (client.region). Views that are private to a person are not part of the file. A view is matched by its name; the main view of an object is matched whatever it is called. |
objects[].permissions |
Per role what may be read, changed and deleted, which fields a role cannot see or change, and the record rules of a role. |
objects[].layout |
The layout of the record page, by field key. Left out when the object uses the built-in layout. |
objects[].duplicate_criteria |
The fields that tell two records apart. |
dashboards |
The dashboards shared with the workspace, matched by name: description, date range and widgets (type, place, settings by object and field key). Private dashboards are not part of the file. With a subset export only dashboards whose charts all use the exported objects come along. A dashboard the file leaves out is deleted only when you also switch off what the file leaves out. |
relations |
Each relation once, by the object and key of both ends. The notes, tasks and files every object has are not listed. |
Contact fields that were created in the older contact settings belong to your contacts and are not part of the file. Fields you add to the contact object on the data model screen are.
API and AI agents
The file is a first-class API object.
| To | REST | Agent tool |
|---|---|---|
| Export | GET /crm/schema (?objects=a,b for a subset) |
get_crm_schema |
| Compare | PUT /crm/schema?dry_run=true |
plan_crm_schema |
| Import | PUT /crm/schema (prune, allow_destructive, async) |
apply_crm_schema |
| Follow a background import | GET /crm/schema/applies/{id} |
get_crm_schema_apply |
The same call exists in the PHP and JavaScript SDKs as schema, importSchema and schemaApply.
Read the API reference for the answers.
Only the owner and administrators can use these, and the plan must include custom objects.
Use it as a template
A subset export is a template: export the objects of a solution (say projects and clients), keep the file with your agency documents, and import it into each new client workspace. Run the comparison first to see exactly what it will set up. Importing a template never touches records the workspace already has.
