Coda Integration - Lead Intelligence Engine

Overview

The Lead Intelligence Engine automatically saves qualified leads to a Coda table. This guide covers Coda setup, column mapping, duplicate detection, and troubleshooting.

Coda Setup

Prerequisites

A Coda account (free tier works)
A Coda document to serve as your CRM
A table within that document for lead storage

Step 1: Create Your CRM Table

Create a New Coda Document

Go to coda.io
Click “New doc”
Name it “Lead Intelligence CRM” (or your preferred name)

Add a Table

Type /table to insert a table
Name it “Leads” or “Business Leads”

Configure Columns

Add these columns (exact names matter):

Column Name	Type	Description
Business URL	Text	Website URL (used for duplicate detection)
Business Name	Text	Extracted business name
Business Type	Text	What the business does
Primary Service	Text	Recommended primary service
Secondary Service	Text	Optional secondary service
Fit Score	Number	Score from 0-100
Reasoning	Text	AI’s reasoning for the recommendation
Outreach Angle	Text	Suggested outreach strategy

You can add extra columns for manual notes, follow-up dates, or status tracking. The engine only writes to the 8 columns listed above.

Step 2: Get Your Coda API Token

Navigate to Account Settings

Go to coda.io/account

Generate API Token

Scroll to API Settings
Click Generate API token
Give it a name like “Lead Engine”
Copy the token (you won’t see it again)

Add to .env

CODA_API_TOKEN=your_token_here

API tokens grant full access to your Coda workspace. Keep them secret and never commit to version control.

Step 3: Get Your Document ID

Open Your CRM Document

Navigate to your Lead Intelligence CRM document in Coda

Extract Doc ID from URL

Your URL looks like:

https://coda.io/d/_dABCDEFGHI/Lead-Intelligence-CRM

The Doc ID is the part after /d/: _dABCDEFGHI

Add to .env

CODA_DOC_ID=_dABCDEFGHI

Step 4: Get Your Table ID

Method 1: Copy Link
Method 2: API Explorer

Right-click on your table
Select “Copy table link”

Paste the URL somewhere:

https://coda.io/d/_dABCDEFGHI/Lead-Intelligence-CRM#_table_grid-abc123xyz

The Table ID is after #_table_: grid-abc123xyz
Add to .env:
```
CODA_TABLE_ID=grid-abc123xyz
```

Go to coda.io/developers/apis
Navigate to “List tables” endpoint
Enter your CODA_DOC_ID
Click “Try it”
Find your table in the response and copy its id field
Add to .env:
```
CODA_TABLE_ID=grid-abc123xyz
```

Verification

Your .env should now have:

.env

CODA_API_TOKEN=a1b2c3d4-e5f6-7890-abcd-ef1234567890
CODA_DOC_ID=_dABCDEFGHI
CODA_TABLE_ID=grid-abc123xyz

Test the connection:

python -c "from coda_client import CodaClient; c = CodaClient(); print('Connection successful!' if c._get_columns() else 'Connection failed')"

Column Mapping

The engine maps evaluation results to Coda columns as follows:

Automatic Mapping

# From evaluator output -> To Coda column
{
  "url" → "Business URL",
  "business_name" → "Business Name",
  "business_type" → "Business Type",
  "primary_service" → "Primary Service",
  "secondary_service" → "Secondary Service",
  "fit_score" → "Fit Score",
  "reasoning" → "Reasoning",
  "outreach_angle" → "Outreach Angle"
}

Column Name Requirements

Column names are case-sensitive and must match exactly. “Business URL” works, “business url” does not.

Correct:

✓ Business URL
✓ Business Name
✓ Primary Service

Incorrect:

✗ business url
✗ BusinessName
✗ Primary_Service

Missing Columns

If a column doesn’t exist in your table, the engine skips it with a warning:

Note: Skipping columns missing in Coda table: Secondary Service

The insert still succeeds with the available columns.

Custom Column Order

Column order doesn’t matter. The engine matches by name, not position:

# This works fine:
[Fit Score] [Business Name] [Business URL] [Primary Service] ...

Additional Columns

You can add extra columns for manual tracking:

Status (Select): “New”, “Contacted”, “Converted”, “Rejected”
Assigned To (Person): Sales team member
Follow-up Date (Date): When to reach out
Notes (Text): Manual observations
Created (Date): Use Coda’s formula =thisRow.CreatedDate()

The engine only writes to its 8 mapped columns and leaves others untouched.

Duplicate Detection Logic

The engine prevents duplicate entries using the Business URL field.

How It Works

Query Coda

Before inserting, the engine queries Coda:

query = f'Business URL:"{url}"'

Check for Matches

If any row has a matching Business URL, the engine skips insertion:

SKIPPED CRM INSERTION: https://example.com
Reason: Duplicate found in CRM

Insert if Unique

If no match found, the engine inserts a new row.

Key Behaviors

URL Normalization

The engine performs exact string matching without normalization. These are treated as different:

https://example.com
https://www.example.com
http://example.com
https://example.com/

Manually clean URLs before analysis if you want strict deduplication (e.g., always include www.).

Case Sensitivity

URLs are case-sensitive:

https://Example.com ≠ https://example.com

Most URLs are lowercase, but edge cases exist.

Query Limitations

If the Coda API query fails (network error, auth issue), the engine:

Logs a warning: Duplicate check failed (Coda API): <error>
Assumes no duplicate exists (fail-safe to prevent data loss)
Proceeds with insertion

This means occasional duplicates may slip through if Coda is unreachable.

Multiple Tables

Duplicate detection only checks the configured CODA_TABLE_ID. If you use multiple tables, each operates independently.

Preventing False Duplicates

If you’re getting “duplicate” warnings for new URLs:

Check for trailing slashes: https://example.com/ vs https://example.com
Verify protocol consistency: http:// vs https://
Check for URL redirects (e.g., example.com → www.example.com)

Forcing Re-analysis

To re-analyze a URL already in Coda:

Delete the row in Coda
Run the analysis again

Or:

Temporarily change the URL in Coda to something else
Run the analysis (will insert new row)
Delete the old row

There is no “force” flag to bypass duplicate detection. This is intentional to prevent accidental CRM pollution.

Data Type Handling

The engine sanitizes data before sending to Coda:

Null Values

# If evaluation returns null for secondary_service:
{"secondary_service": null}

# Converted to empty string before Coda insertion:
{"secondary_service": ""}

Coda API rejects null values, so the engine converts them to empty strings.

Special Characters

No escaping needed. The engine handles:

Quotes: Business name: "Joe's Plumbing"
Newlines: Multi-line reasoning text
Emojis: 🚀 Awesome Business
Unicode: Café François

Number Fields

The Fit Score column should be type Number in Coda:

{"fit_score": 85}  # Inserted as number, not string "85"

If you accidentally set it as Text, Coda will still accept it but won’t allow numeric sorting/filtering.

Error Handling

Common Coda Errors

”Coda configuration is incomplete in .env”

Cause: Missing one or more Coda variables Fix: Verify all three variables exist in .env:

CODA_API_TOKEN=...
CODA_DOC_ID=...
CODA_TABLE_ID=...

“401 Unauthorized”

Cause: Invalid or expired API token Fix:

Go to coda.io/account
Regenerate API token
Update .env with new token

”404 Not Found”

Cause: Incorrect Doc ID or Table ID Fix:

Verify Doc ID by checking document URL
Verify Table ID using “Copy table link” method
Ensure no extra spaces in .env

”403 Forbidden”

Cause: API token lacks write permissions Fix:

Regenerate token with full permissions
Ensure you own the document (not just a viewer/commenter)

“Column ‘X’ not found”

Cause: Column name mismatch (case-sensitive) Fix:

Check exact column names in Coda
Rename to match required names:
- Business URL (not Business url or URL)
- Business Name (not Name or Company)
- etc.

Network Issues

Timeout Errors

Coda API Error: ReadTimeout | Details: Request timed out after 15s

Fix:

Check internet connection
Verify Coda status at status.coda.io

Increase timeout in coda_client.py:111:

response = requests.post(url, headers=self._get_headers(), json=row_payload, timeout=30)  # Increased from 15

Rate Limiting

Coda API Error: 429 Too Many Requests

Coda’s API has rate limits:

Free tier: ~10 requests/second
Team/Enterprise: Higher limits

Fix:

Add delays between batch operations
Reduce parallel analysis (if running multiple bots)
Upgrade Coda plan for higher limits

Troubleshooting

Testing Coda Connection

Run the built-in test:

python coda_client.py

Expected output:

Coda client loaded. Set .env variables to test.

For a full insertion test, uncomment lines 135-139 in coda_client.py:

coda_client.py

if __name__ == "__main__":
    client = CodaClient()
    mock_data = {
        "url": "https://test.com",
        "business_name": "Test Business Name",
        "business_type": "Test Business Type",
        "primary_service": "Foundation Package",
        "secondary_service": None,
        "fit_score": 90,
        "reasoning": "Test reasoning",
        "outreach_angle": "Test hook"
    }
    try:
        res = client.insert_row(mock_data)
        print("Insert successful:", res)
    except Exception as e:
        print("Error:", e)

Run:

python coda_client.py

Should insert a test row into your table.

Viewing API Logs

Enable detailed logging:

import logging
logging.basicConfig(level=logging.DEBUG)

Add to top of coda_client.py temporarily for verbose output.

Inspecting API Responses

Modify coda_client.py:111 to print raw responses:

response = requests.post(url, headers=self._get_headers(), json=row_payload, timeout=15)
print("Response:", response.status_code, response.text)  # Add this line
response.raise_for_status()

Common Column Mapping Issues

Column names with extra spaces

# Wrong in Coda:
"Business Name "  (trailing space)
" Business Name"  (leading space)

# Correct:
"Business Name"

Fix: Remove extra spaces in Coda column headers

Column type mismatch

If Fit Score is set as Text instead of Number, you’ll get:

Coda API Error: Invalid value for column 'Fit Score'

Fix: Change column type to Number in Coda

Required columns in Coda

If you set Coda columns as “Required” but the engine sends empty values:

Coda API Error: Required column 'Secondary Service' cannot be empty

Fix: Make the column optional in Coda, or ensure the AI always provides a value

Advanced Configuration

Custom Column Mapping

If you want different column names, edit coda_client.py:78-87:

coda_client.py

desired_cells = [
    {"column": "URL", "value": evaluation_data.get("url")},  # Changed from "Business URL"
    {"column": "Company Name", "value": evaluation_data.get("business_name")},  # Changed
    # ... customize others ...
]

This requires code changes. Prefer using the standard column names to avoid maintenance overhead.

Multiple CRM Tables

To write to different tables based on conditions:

# In core.py, customize insert_row call:
if result.get("fit_score") >= 80:
    self.coda.table_id = os.getenv("CODA_TABLE_ID_HIGH_PRIORITY")
else:
    self.coda.table_id = os.getenv("CODA_TABLE_ID_STANDARD")

self.coda.insert_row(result)

Requires setting up multiple table IDs in .env.

Webhook Notifications

Use Coda automations to trigger webhooks when new rows are added:

In Coda, click Automations (top right)
Create rule: “When rows are added to Leads”
Action: “Send webhook” → Zapier, Make.com, or custom endpoint

This enables:

Slack notifications for new leads
Auto-adding to other CRMs (HubSpot, Salesforce)
Triggering email campaigns

Next Steps

CLI Usage

Start analyzing leads via command line

Telegram Bot

Use the bot interface for easier lead generation

Get Started

Core Concepts

Usage Guides

Knowledge Base

​Overview

​Coda Setup

​Prerequisites

​Step 1: Create Your CRM Table

​Step 2: Get Your Coda API Token

​Step 3: Get Your Document ID

​Step 4: Get Your Table ID

​Verification

​Column Mapping

​Automatic Mapping

​Column Name Requirements

​Missing Columns

​Custom Column Order

​Additional Columns

​Duplicate Detection Logic

​How It Works

​Key Behaviors

​Preventing False Duplicates

​Forcing Re-analysis

​Data Type Handling

​Null Values

​Special Characters

​Number Fields

​Error Handling

​Common Coda Errors

​”Coda configuration is incomplete in .env”

​“401 Unauthorized”

​”404 Not Found”

​”403 Forbidden”

​“Column ‘X’ not found”

​Network Issues

​Timeout Errors

​Rate Limiting

​Troubleshooting

​Testing Coda Connection

​Viewing API Logs

​Inspecting API Responses

​Common Column Mapping Issues

​Advanced Configuration

​Custom Column Mapping

​Multiple CRM Tables

​Webhook Notifications

​Next Steps

CLI Usage

Telegram Bot

Build docs developers (and LLMs) love

Overview

Coda Setup

Prerequisites

Step 1: Create Your CRM Table

Step 2: Get Your Coda API Token

Step 3: Get Your Document ID

Step 4: Get Your Table ID

Verification

Column Mapping

Automatic Mapping

Column Name Requirements

Missing Columns

Custom Column Order

Additional Columns

Duplicate Detection Logic

How It Works

Key Behaviors

Preventing False Duplicates

Forcing Re-analysis

Data Type Handling

Null Values

Special Characters

Number Fields

Error Handling

Common Coda Errors

”Coda configuration is incomplete in .env”

“401 Unauthorized”

”404 Not Found”

”403 Forbidden”

“Column ‘X’ not found”

Network Issues

Timeout Errors

Rate Limiting

Troubleshooting

Testing Coda Connection

Viewing API Logs

Inspecting API Responses

Common Column Mapping Issues

Advanced Configuration

Custom Column Mapping

Multiple CRM Tables

Webhook Notifications

Next Steps