Peaka

@Peaka list all tables and columns in the 'sales_analytics' project so I can see what data is available to query

The sales_analytics project contains 3 catalogs. In the 'crm' catalog, you'll find tables like 'customers' (columns: customer_id, name, email, created_at), 'deals' (deal_id, customer_id, amount, stage, close_date), and 'interactions' (interaction_id, deal_id, type, timestamp). The 'marketing' catalog includes 'campaigns' and 'conversions' tables. The 'finance' catalog has 'invoices' and 'payments'. This illustrative example shows how Peaka surfaces the schema structure you need before writing SQL queries.

This read-heavy example demonstrates schema discovery — essential before querying unfamiliar data sources. Requires read access to the target project. The actual catalog and table names depend on what data connections your Peaka project has configured; this example assumes a typical CRM-plus-marketing setup.

@Peaka execute a query in project 'sales_analytics': SELECT customer_id, SUM(amount) as total FROM deals WHERE stage = 'closed_won' GROUP BY customer_id ORDER BY total DESC LIMIT 10

Query executed successfully. Top 10 customers by closed deal value: customer_id 'C-1047' ($487,200), 'C-0923' ($412,500), 'C-1138' ($389,000), continuing down to 'C-0654' ($201,300). This illustrative result shows how Peaka returns aggregated data from your connected sources in a structured format, ready for AI analysis or follow-up questions about trends and outliers.

This synthesis example pairs SQL execution with AI reasoning potential — you can ask follow-up questions about the results. Queries are read-only and capped at 1,000 rows. Execution time depends on the underlying data source's performance; complex joins across federated sources may be slow.

@Peaka create a new project called 'Q1_Revenue_Analysis' in the 'finance_team' workspace with description 'Quarterly revenue breakdown by region and product line'

Project created successfully. Project ID: 'prj_8x9k2m4n', Name: 'Q1_Revenue_Analysis', Workspace: 'finance_team'. The project is now available for connecting data sources and running queries. This illustrative confirmation shows the project ID you'll reference in subsequent data operations and the workspace context where team members can access it.

This write example demonstrates workspace organization — useful when spinning up ad-hoc analysis environments. Requires workspace-level create permissions. Creating a project does not automatically provision data connections; you'll configure those separately in the Peaka UI or via additional API calls.

When Peaka wins for cross-platform data pulls at small scale

A 6-person ops team needs to join Stripe revenue data with Salesforce pipeline data for weekly board decks. Peaka's Execute Query tool lets you write SQL that spans both sources without building a data warehouse. You create one project per reporting domain, run SELECT queries capped at 1,000 rows, and pull results into a shared Switchy thread. This works if your queries stay under the row limit and you're comfortable with API key auth (no SSO). If you're pulling more than 1,000 rows per query or need real-time refresh under 5 minutes, you'll hit Peaka's execution ceiling and should look at a dedicated ETL tool. For weekly or daily reporting at sub-1k row scale, Peaka keeps your team out of CSV hell without hiring a data engineer.

Peaka for federated search across support tools

A 4-person support team uses Zendesk for tickets, Notion for internal docs, and Airtable for product FAQs. Peaka's List Data Resources and Execute Query tools let you catalog all three sources in one project, then run SQL joins to answer questions like 'show me all tickets tagged refund where the FAQ exists in Airtable'. You set up the project once, share the workspace ID in Switchy, and any teammate can query without learning three different APIs. The trade-off: Peaka's 1,000-row limit means this works for targeted lookups, not bulk exports. If you're trying to pull every ticket from the last year, you'll need batch tooling. For day-to-day 'find me the 10 relevant docs' queries, Peaka beats writing custom scripts or paying for a full search platform.

When Peaka handles monthly close for distributed finance data

A 3-person finance team reconciles Stripe payouts, QuickBooks invoices, and Gusto payroll every month. Peaka's Execute Query tool lets you write SQL that joins all three sources in one pass, returning discrepancies as a 200-row result set. You create a 'monthly-close' project, run the reconciliation query, and paste results into a Switchy thread for review. This works because the row count is predictable and the query runs once a month. If you need sub-hour latency or you're processing 5,000+ transactions per close, Peaka's execution model (read-only, 1k row cap) won't scale. For monthly or quarterly close workflows where the output fits in a spreadsheet, Peaka eliminates the export-import-vlookup loop without standing up Airflow.