CRM & Notion Field Filler — Enrich Records You Already Have

Your CRM is full of half-finished records. A company name and nothing else. A contact with no phone. A lead with no website.

This Actor finds those gaps and fills them — only the blanks, never the values a person typed in by hand.

Point it at the app your records live in (Notion, HubSpot, Supabase, Airtable, any MCP connector), point it at an Apify dataset that has the missing values, tell it which field joins the two, and it patches the difference.

What makes it safe to run on a live CRM

• Blank-only by default. A field holding a value is left alone. overwrite exists, and is off.
• Minimal patches. It sends only the fields it actually filled, not the whole record.
• No LLM. Deterministic joins on a key you choose. Nothing is inferred, guessed, or hallucinated into your database.
• dryRun reports every record and field it would touch, for free, before anything is written.
• One rejected record never aborts the rest of the run.

How it works

1. Connect your app under Integrations in Apify Console.
2. Run in list-tools mode to see the connector's tool names.
3. Run in fill mode.

{
  "destination": "<your connector>",
  "mode": "fill",
  "readTool": "query_database",
  "readToolArgs": { "database_id": "abc123" },
  "updateTool": "update_page",
  "idField": "id",
  "idArgName": "page_id",
  "datasetId": "<dataset with the missing values>",
  "matchKey": "email",
  "fillFields": ["phone", "website", "owner"]
}

Credentials never reach this Actor. It talks to your app through Apify's MCP proxy, which holds the real tokens.

Examples

Fill missing phone numbers on CRM contacts — matchKey: "email", fillFields: ["phone"], source dataset from any contact-enrichment Actor.

Add owner names to a contractor list — matchKey: "licenseNumber", fillFields: ["owner", "phone"], source from a licence-registry Actor.

Backfill company websites in Notion — matchKey: "domain", fillFields: ["website", "industry"].

Input

Field	Required	Description
destination	yes	MCP connector holding your records
mode		fill (default) or list-tools
readTool	for fill	Tool that lists your existing records
readToolArgs		JSON args for the read tool
updateTool	for fill	Tool that patches one record
updateToolExtraArgs		Constant args merged into every update
idField		Where the record ID lives in the read result. Default id
idArgName		What the update tool calls that ID. Default id
datasetId	for fill	Apify dataset supplying the values
matchKey	for fill	Field joining both sides. Case/space insensitive
sourceMatchKey		Only if the source names the join field differently
fillFields	for fill	Destination fields to complete
fieldMap		Only if a source field is named differently or nested
overwrite		Also replace non-blank values. Default off
maxItems		Cap on records updated per run. Default 1000
dryRun		Plan only, write nothing

Output

One row per record handled:

{ "updated": true, "key": "a@example.com", "id": "page_123",
  "filled": ["phone"], "patch": { "phone": "480-555-0100" } }

Failures are recorded rather than dropped:

{ "updated": false, "key": "b@example.com", "id": "page_124", "error": "Tool \"update_page\" returned an error: ..." }

What counts as blank

null, missing, "", whitespace, and []. Note that 0 and false are real values and are not treated as gaps.

Pricing

$5 per 1,000 records updated. Records that were already complete, records with no match in the source, and dry runs are all free. You pay when a record actually gained data.

FAQ

Does this see my Notion/HubSpot token? No. Apify's MCP proxy holds credentials and attaches them to outbound calls.

Will it overwrite my data? Not unless you set overwrite: true. By default it only writes into empty fields.

What if a record has no match in the source? It's skipped and counted as unmatched. Nothing is written.

Can an AI agent call this? Yes — it's exposed through the Apify MCP server, and list-tools then fill is a natural two-step.