Turn a client's connected feed into questions you can work with through Catchr MCP. Discover its available columns, choose the data relevant to your report and ask ChatGPT to summarize records or investigate values before your next client call.
For freelancers and agencies working with custom feeds, each analysis starts with the fields that feed actually contains.
“Compare paid media performance across my client accounts and flag the biggest changes.”
JSON / CSV / XML
Google Ads
GA4
Used by the world's leading companies
JSON, CSV and XML feeds have their own structures. Discover the available fields first, then select the columns and calculations that fit your reporting question. There is no fixed set of performance metrics for this source.
Explore scalar fields such as text and numbers, including fields flattened from nested objects. Use the discovered columns to decide which values to group, compare or include in a client summary.
Arrays are excluded from field discovery. Information inside an array should not be assumed available as a queryable column.
Use Source URL to identify the resolved feed address after date placeholders are applied, and Response Format to see the format used to parse the response.
Row Index gives the sequential position of a record in the fetched payload. It can help reference rows during a review, but should not be treated as a persistent record identifier across requests.
For feeds with usable date data, explore reporting dimensions such as Date, Year month, Year week and Year Quarter. Confirm the relevant date field before comparing periods.
Date placeholders can also be applied to the feed URL, request headers and POST body. The records returned depend on the configured request and the feed’s response; requesting a period does not establish that the feed contains records for it.
A focused feed review, a reusable report and a question through Catchr MCP serve different parts of client work. Start with the available columns and choose the method that matches the answer you need.
Use a feed review to understand the available data, a report to present a consistent view and Catchr MCP to investigate specific client questions. Confirm field meanings and date coverage before drawing conclusions from any of these methods.
Connect your remote feed URL in Catchr. Configure the response format, HTTP method and record path, plus request headers or a request body when the feed requires them. Date placeholders can be used in the URL, headers and POST body.
Choose ChatGPT or Claude through Catchr MCP, discover the connected feed’s columns and select the fields for your first question.
Use these prompts to explore a client feed, review its records or prepare a reporting summary. Replace the bracketed values with your connected feed and choices from its discovered columns.
Using Catchr, discover the available columns in [connected feed]. List their names and available types, distinguish feed columns from Source URL, Response Format and Row Index, and ask me which columns to use for [client reporting question]. Do not assume fields that are absent from discovery.
Copy promptUsing Catchr, discover the fields in [connected feed] and confirm that [selected numeric column] supports a numeric comparison. Return records where its value exceeds [threshold], showing [selected reference columns] and Row Index. Sort by the selected numeric value and report how many matching records are in the query results.
Copy promptUsing Catchr, discover the columns in [connected feed] and confirm that [selected date column] and [selected numeric column] support a period comparison. Compare [first date range] with [second date range] using [chosen aggregation], grouped by [selected category column]. Check the returned date coverage and flag missing periods before describing differences.
Copy promptUsing Catchr, discover the fields in [connected feed] and review [selected reporting columns] for missing values. Check repeated values in [selected reference column] without assuming they are errors. Summarize the findings, show affected records with Row Index, and explain which gaps could affect [client report].
Copy promptYou can work with the feed’s discovered scalar columns and fields flattened from nested objects. Select the relevant columns to summarize values, review records or compare groups. Available metrics depend on the feed; this connector does not provide a fixed set.
Start with the remote feed URL. Configure its response format, HTTP method and record path, along with any required headers or request body. Then connect your AI destination through Catchr MCP and discover the fields before asking reporting questions.
Arrays are excluded from dynamic field discovery. Scalar fields and nested scalar objects can become queryable columns, but values inside arrays are not inferred. Check the discovered field list before including a column in your report or prompt.
Date placeholders can be substituted into the source URL, request headers and POST body. The feed determines which records that request returns. Check date coverage and usable date fields before asking for weekly, monthly or period-over-period summaries.
Connect your sources to Catchr, add the MCP server to your AI assistant, and ask about the work that matters now.
