How to Import Airtable into DenchClaw

Importing your Airtable data into DenchClaw takes about 15 minutes for most bases. Export your tables as CSV, map the columns to DenchClaw fields, and use the built-in importer or browser agent to pull everything across. This guide walks through every step — including how to handle linked records, attachments, and multi-select fields that trip people up.

What is DenchClaw? — it's a local-first, open-source AI CRM that stores everything in DuckDB on your machine. No cloud vendor, no per-seat fees, no data leaving your laptop.

Why migrate from Airtable to DenchClaw?#

Airtable is a great spreadsheet-database hybrid. But when your data is your business — your contacts, deals, client relationships — you probably don't want it living in someone else's cloud under their pricing schedule. Airtable's free tier caps rows, paid plans get expensive fast, and you have no control over the schema at the database level.

DenchClaw gives you:

• Full ownership — DuckDB file on your disk
• SQL access to everything, any time
• AI-powered relationship intelligence
• Browser agent that's already logged in everywhere you are
• MIT licensed, self-hostable

If that sounds like what you want, let's move your data.

Step 1: Export your Airtable base as CSV#

1. Open your Airtable base
2. Click the Grid view for the table you want to export
3. Click the table name dropdown → Download CSV
4. Repeat for each table you want to migrate

Tip: If you have linked records across tables, export both tables. You'll reconnect them using relation fields in DenchClaw.

For bases with attachments (file uploads, images): Airtable's CSV export replaces attachment cells with URLs. You can either download those files separately or just keep the URLs as-is in DenchClaw.

Step 2: Understand your schema#

Before importing, map out what you have:

Write this mapping down — you'll need it when setting up your DenchClaw object.

Step 3: Install DenchClaw#

npx denchclaw

This spins up on port 19001 with the web frontend at localhost:3100. First run takes ~30 seconds to initialize the DuckDB database.

Step 4: Create your object in DenchClaw#

Every data type in DenchClaw is an "object" — think of it like a table in Airtable. You define it in a Object.yaml file or through the UI.

Here's how to do it through the UI:

1. Open localhost:3100
2. Click + New Object in the sidebar
3. Name it (e.g., "Contacts", "Projects", "Leads")
4. Add fields matching your schema mapping from Step 2

For each field, select the right type. Pay attention to select fields — you'll need to pre-populate the options to match your Airtable choices, otherwise the importer will create them automatically.

Step 5: Import the CSV#

DenchClaw's importer accepts CSV directly:

1. Open your object in the Table view
2. Click Import → CSV
3. Upload your exported Airtable CSV
4. Map each CSV column to a DenchClaw field
5. Preview the first 5 rows
6. Click Import

For a 10,000-row table, expect about 10-20 seconds.

Multi-select fields: DenchClaw expects multi-select values as comma-separated strings in the CSV cell (e.g., "tag1, tag2, tag3"). Airtable exports them exactly this way — you're fine.

Step 6: Handle linked records (relation fields)#

This is where Airtable migrations usually get messy. Here's how to do it cleanly.

Say you have a Contacts table linked to a Companies table.

1. Import Companies first — note that DenchClaw assigns each entry a numeric ID
2. Import Contacts without the relation column for now
3. Use DuckDB to create the links:

-- Connect contacts to companies by name
UPDATE entry_fields ef
SET value = c.id
FROM entries e
JOIN v_contacts vc ON vc.id = e.id
JOIN v_companies c ON c.name = vc.company_name
WHERE ef.entry_id = e.id
AND ef.field_key = 'company_id';

Replace company_name and company_id with your actual field names. This pattern works for any many-to-one relation.

Step 7: Validate your import#

Run a quick sanity check in DuckDB:

-- Check row counts match
SELECT COUNT(*) FROM v_contacts;
 
-- Check for missing required fields
SELECT id, name FROM v_contacts WHERE email IS NULL;
 
-- Check multi-select fields parsed correctly
SELECT id, tags FROM v_contacts WHERE tags IS NOT NULL LIMIT 10;

Open the DenchClaw UI and spot-check a few records against your original Airtable base.

Step 8 (Optional): Use the browser agent for Airtable API imports#

If you have a large Airtable base (50,000+ rows) or need to pull data from Airtable's API rather than CSV, DenchClaw's browser agent can help.

The browser agent copies your Chrome profile — so it's already logged into Airtable. You can write a skill that:

1. Opens Airtable's API endpoint for your base
2. Paginates through all records
3. Writes directly to DuckDB

This is especially useful for automating ongoing syncs before you fully cut over.

See browser automation for CRM for the full pattern.

What to do after importing#

Once your data is in DenchClaw, you have a few immediate wins available:

• Set up views: Table, Kanban, Calendar, Timeline, Gallery — configure them in the UI
• Add AI enrichment: Use action fields to trigger enrichment scripts on any entry
• Create entry documents: Each contact/company/deal gets a markdown file for notes
• Connect channels: Hook up Telegram, WhatsApp, or iMessage so DenchClaw knows about your conversations

Also see how to import from Notion and how to import from HubSpot if you're consolidating from multiple tools.

FAQ#

Does DenchClaw support Airtable's formula fields? Not directly — but DuckDB is a full SQL engine. Recreate any formula as a computed column in a DuckDB view. That view is then accessible in DenchClaw as a read-only field.

What happens to Airtable automations? Automations don't migrate. DenchClaw uses action fields (buttons that execute server-side scripts) and skills (markdown plugins) for automation. You'll need to recreate these, but they're more powerful.

Can I keep Airtable and DenchClaw in sync during migration? Yes — use the browser agent to pull from Airtable's API on a schedule while you're transitioning. Once you're confident, cut over fully.

What about Airtable's gallery and calendar views? DenchClaw has both Gallery and Calendar views built in. Map your date fields and image attachment fields after importing.

I have 200+ fields in my base. Is there a faster way? Write a script that reads your CSV headers and generates the DenchClaw Object.yaml field definitions automatically. A GPT-assisted script can do this in under a minute.

Ready to try DenchClaw? Install in one command: npx denchclaw. Full setup guide →