Skip to content

Collections

A collection is a notebook. Its requests sit in numbered steps, the ones outside a section first and then each section in turn. Requests of any type can share one collection.

The New collection dialog (the + in the sidebar, or a tile on Home) has three choices:

  • Empty collection: a name, an icon, and an optional Base URL that Codestamp saves as the collection variable {{baseUrl}}.
  • Import a file: pick a file, then check the preview on the Import page. See Import and export.
  • Join a shared collection: paste a Git repository address. See Teams.

A new, empty collection shows an “Add your first request” panel with HTTP, WebSocket and Paste cURL.

  • One request card is open at a time. The others show a summary.
  • Codestamp autosaves 600 ms after you stop typing, and Ctrl+S saves at once.
  • To reorder, drag a request or a section, press Alt+↑ or Alt+↓, or pick a section in the card header to move the request into it.
  • Deleting a section keeps the requests inside it. Deleting a collection deletes everything in it, after a confirmation.
  • Duplicate copies a request, a section or a whole collection. A copied collection brings its auth, sections, requests and collection variables, and Codestamp re-encrypts secret variables for the copy.
  • Each collection has an emoji icon. Codestamp picks one from the name, and you can change it in the picker.
  • Quick search (Ctrl+Space) finds requests, collections and Settings pages. Ctrl+F searches inside an editor.

A workspace keeps its collections, environments, Globals, flows and cookies apart from the others. Each has its own icon: a letter, a company logo, an image or an emoji. Switch workspaces from the title bar.

A collection or request you delete stays in Recently deleted for 30 days. A banner offers Undo right after. The Recently deleted page restores an item with its sections, examples, Docs, variables, environments, team link and Orchestration steps, or deletes it for good.

A restored request goes back to its section if that section still exists. If its collection is gone too, Codestamp restores it into a new collection with the same name.

Run all, in the collection header, runs every HTTP and GraphQL request in notebook order.

  • A request passes when a response arrives, its scripts do not fail, no test fails, and either it has tests or the status is below 400.
  • WebSocket requests are skipped.
  • You can stop at the first failure and add a pause between requests of 0, 250 or 1000 ms. You can stop the run at any time.
  • Run only this runs one request. You can export the result as a JSON report without bodies.
  • A run does not write to History.

For a data run, give Run all a CSV file (the first row holds the column names) or a JSON file (an array of objects), up to 5 MB and 5,000 rows. It runs once per row, each column becomes a {{variable}} that overrides every scope, and scripts can read pm.iterationData, pm.info.iteration and pm.info.iterationCount. Without a data file you can set the number of iterations from 1 to 100.

An Orchestration flow runs requests one after another between a Start and an End node, drawn as a step map from left to right.

  • Each step calls an existing HTTP or GraphQL request from any collection. A step is a reference, so a change to the request changes the flow. WebSocket, MCP and gRPC requests cannot be picked.
  • Pass on takes a value from a response body (a JSON path such as data.items[0].id) or a header and hands it to later steps as {{name}}. During a run these values win over the environment, and Codestamp stores them nowhere.
  • Checks need no script: status, body, header and response time. The request’s own pm.* script still runs, and its pm.test results count.
  • A step passes when a response arrives, its script and Pass on do not fail, and every check and test passes. With no checks and no tests, it passes when the status is below 400.
  • If a step fails, the flow stops (the default) or continues with the next step.
  • A flow can have 100 steps, 20 checks and 20 values per step, and each value can be up to 8 KB. A flow run does not write to History. Flows do not sync through Collaboration yet and cannot run through the CLI.

History is per collection. The History button in the collection header lists the requests sent from it, with search, a status filter, Open, Save to collection, and Clear history.

Docs, under Tools, writes documentation for each collection. You can add a Markdown description to a collection, a section and a request, up to 20,000 characters, and Docs shows it next to the URL, auth, parameters, headers, body and saved examples. Credential values typed directly into a request are masked there, and {{variable}} references stay visible. Save as Markdown writes one .docs.md file.