#What the importer does

The importer lives at /dashboard/import ("Import Data"). You pick an import type, upload a CSV, map columns to fields, review a preview, and run the import. Everything is scoped to your current tenant, so imported records only ever appear in your own workspace.

There are two import types, chosen with the toggle at the top of the page. Each writes to a different table and shows up in a different part of the app:

#Prepare your CSV

The parser reads the first line as headers, handles quoted values, and understands doubled quotes ("") as an escaped quote inside a field. Values are trimmed and fully blank rows are skipped automatically. Commas are the default delimiter.

Core fields you can map

Each import type exposes a fixed set of core fields. Fields marked required must be mapped, or the import will not start. Rows that are missing a value for a required field are skipped and counted separately.

emailrequired

Contact / lead email. Required for both leads and contacts.

first_namerequired

Given name. Required for both types.

last_namerequired

Family name. Required for both types.

phoneoptional

Phone number, any format (e.g. (555) 123-4567).

companyleads only

Company name. A leads-only core field on the import page.

statusleads only

Lead status such as new. Defaults to new if left blank (database default).

sourceleads only

Lead source such as website or referral; the CSV importer writes it to the lead's source column as entered.

addresscontacts only

Street address for the contact.

#Run an import step by step

1. 1
   Open Import Data

   Go to /dashboard/import and choose Leads or Contacts with the toggle. Switching types clears the current mapping, preview, and result.
2. 2
   Upload your CSV

   Click the file picker and select a .csv file. The file is parsed in your browser — nothing is uploaded to a server at this stage.
3. 3
   Review the auto-mapping

   The importer matches each CSV header to a field by normalizing names (lowercasing, stripping non-alphanumerics) and looking for exact or partial matches. Adjust any dropdown that guessed wrong, and set unwanted columns to "-- Skip --".
4. 4
   Check the preview

   The first 10 rows render in a table so you can confirm the data lines up with your columns. The total row count is shown beneath it.
5. 5
   Import

   Click "Import N Records". Records are inserted in batches of 100 directly into crm_leads or crm_contacts, and an import_jobs record tracks progress.
6. 6
   Read the result

   A summary shows imported, failed, and skipped counts. Any database errors are listed under "View Errors". The run also appears in Import History (last 10 jobs).

#Custom fields

Any custom fields you have defined for the selected object type are loaded automatically and appear under a "Custom Fields" group in each mapping dropdown. The page shows a note like "3 custom fields will be available for mapping" when custom fields exist.

• Custom-field columns are mapped with a cf: prefix internally (for example cf:policy_number).
• Mapped custom values are collected into the record's custom_fields JSONB column, not into a dedicated table column.
• The programmatic leads API (below) folds a lead's company value into custom_fields.company, since crm_leads has no dedicated company column.

#Import history and job records

Every run is recorded in the import_jobs table and the most recent ten are shown in the Import History panel. Each job captures how many records were attempted, how many succeeded, and how many failed.

import_typetext

"leads" or "contacts".

statustext

Progresses from processing to completed (or completed_with_errors when some batches fail).

total_recordsint

Number of records that passed required-field validation and were queued.

processedint

Records successfully inserted.

failedint

Records in batches that hit a database error.

errorsjsonb

Any database error messages captured during the run, or null.

completed_attimestamptz

When the job finished.

#Programmatic import (API)

To load leads without the UI — for example from a script or another system — post a JSON array of leads to /api/v1/import/leads. This endpoint is leads-only and takes JSON, not CSV. It requires an authenticated request that resolves to a tenant (the same auth the app uses).

curl -X POST https://your-workspace.olyron.app/api/v1/import/leads \
  -H "Content-Type: application/json" \
  --cookie "your-session-cookie" \
  -d '{
    "leads": [
      {
        "email": "john@example.com",
        "first_name": "John",
        "last_name": "Doe",
        "phone": "(555) 123-4567",
        "company": "Acme Corp",
        "status": "new",
        "source": "referral",
        "custom_fields": { "policy_number": "A-1042" }
      }
    ]
  }'

Each lead requires email; rows without it are rejected and reported per-index. source (or lead_source) is normalized to a canonical value, company is folded into custom_fields, and the endpoint writes its own import_jobs record. The response reports what happened:

{
  "success": true,
  "imported": 1,
  "failed": 0,
  "errors": null,
  "data": [ { "id": "…", "email": "john@example.com", "status": "new" } ]
}

#Limits and troubleshooting

For very large books, split the file into batches. The importer processes rows in batches of 100, and the leads API and export endpoint are bounded (the LinkedIn importer, for example, caps a request at 20,000 rows; exports return up to 10,000 rows per call).