Sourcetable

Sourcetable is a browser spreadsheet with an AI assistant that is meant to remove the two jobs analysts hate: hunting for the right formula and stitching data sources together by hand.

The assistant writes formulas, fixes broken ones, builds pivot tables, cleans and de-duplicates data, generates templates, writes Python when a formula will not do, and produces charts from a prompt.

Underneath it is a real spreadsheet - familiar grid, formulas, sorting, conditional formatting, data validation, imports and exports - so work can be moved in from Excel rather than rebuilt.

Output side is covered too: charts are generated from plain English, and the analysis layer supports summarisation and insight extraction rather than only cell-level edits.

The models behind it are selectable and openly named - OpenAI, Claude and Llama are all credited on the site - which matters if a team has a policy about which provider sees its data. The marketing relevance sits in the connector catalogue and one specific use case.

The site lists Marketing Campaigns and attribution as a first-class use case, and the connectors page names hundreds of sources including Google Analytics 4, Google Ads, Google Search Console, Facebook Ads, LinkedIn Ads, TikTok Ads, Amazon Ads, Pinterest Ads, Snapchat Ads, Shopify, Stripe, HubSpot, Salesforce Marketing Cloud, Snowflake, BigQuery, Postgres, MySQL, ClickHouse, DuckDB and Supabase.

Teams also connect GA4 and live query databases directly for ad hoc questions instead of waiting on a data team.

That means blended reporting - spend from ad platforms, sessions and conversions from GA4, revenue from Shopify or Stripe - can be assembled and interrogated in one sheet, then turned into a chart or a written summary.

On the analysis side, the AI spreadsheet assistant covers AI research and Deep Research, smart formulas, pivots and visualisations, so a question like which channel produced the cheapest qualified trial last quarter can be asked in English rather than written as SQL. Pricing was readable on sourcetable.com/pricing on 24 September 2026 and is credits-based per user.

Regular is free with 50 credits a month, resetting monthly, and includes the AI spreadsheet assistant, AI analysis and summarisation, smart formulas, pivots, charts, magic-autofill, data generation and AI research, Excel template generation, data cleaning and transforms, plus unlimited access to data connectors.

Pro is $29 per user per month with 2,000 credits and adds Deep Research, premium AI models, the Python and data science toolkit, expanded chat memory and context, and a SQL editor and query builder.

Max is $100 per user per month with 6,000 credits, the highest model intelligence, advanced agent capabilities, maximum memory and context and unlimited AI messages subject to abuse guardrails.

Extra credits can be bought as one-time packs from $10 to $200 that do not expire - 2,500 Pro credits work out at $0.02 each and 3,333 Max credits at $0.015 - and education discounts are offered for university emails.

Two limits are worth pricing in before a team subscribes.

First, credits rather than seats are the real ceiling, so heavy iteration eats through 50 free or 2,000 Pro credits quickly and the one-time packs become part of the budget.

Second, this is a spreadsheet, not a governed BI layer: it does not replace scheduled dashboards, row-level permissions or a semantic layer, and Sourcetable itself positions the product around analysis inside a workbook.

Enterprise security is documented - SOC 2 Type 2, HIPAA and PCI, private user keys, encryption at rest and escrow KMS - which is the part that makes pulling warehouse and payment data into a sheet acceptable to a security reviewer.