Vegha for VS Code

A local-first API client inside VS Code, powered by the same engine as the Vegha desktop app. Collections are plain-text Bruno-style .bru files in your workspace, so they diff, review and merge in git like the rest of your code.

Documentation · Getting started · Vegha desktop app · Report an issue

Features

• Visual request editor for .bru files: method, URL, query/path params, headers, every body type (JSON, XML, text, GraphQL, form, multipart, binary file) and auth.
  • The file stays the source of truth. Undo/redo, dirty state, save and git diff all work as in any text file.
  • Paste a cURL command into the URL bar to import it.
• Every auth type: Bearer, Basic, API key, OAuth 2.0 (client credentials, password, and authorization code + PKCE through your browser), OAuth 1.0, AWS SigV4, Digest, NTLM, WSSE, and mTLS client certificates.
• Environments: collection and workspace environments, {{variable}} hovers, and an active-environment picker in the status bar.
  • secret:// references resolve through your configured secret managers (Azure Key Vault, AWS Secrets Manager).
• Scripts and scriptless tests: pre-request and post-response JavaScript (bru / req / res), test() blocks, scriptless assertions and response captures — pick a value from the last response instead of typing its res.body… path.
• Request chaining: in the Chain tab, choose requests to run first from a dropdown, set when each runs (always, only when a variable is empty, only when its last run is older than N minutes, or when a variable matches a value), and drag to reorder. Each card shows what the linked request captures, and {{variables}} a linked request sets are recognised in this one.
• Response viewer:
  • Pretty, raw and HTML preview of the body, plus image previews.
  • Headers, cookies, test results and console output.
  • A timing breakdown, the variables each request set, and the raw request that was sent.
• Collection runner: run a collection or folder with iterations, CSV/JSON data files (one iteration per row), delays and stop-on-failure, carrying variables between requests or running in parallel.
  • Live results with per-request detail.
  • Export the run as HTML, JUnit or JSON.
• Test Explorer: requests appear in VS Code's Testing view. Each assertion and test() is its own test, with the failure message shown inline.
• Import and export:
  • Import Postman, Insomnia, OpenAPI/Swagger, WSDL/SoapUI and Bruno collections, plus Postman environments.
  • Export to Postman, OpenAPI 3, Bruno and Insomnia. Secret values are never exported.
• Cookies view of the shared jar, with delete and clear.
• Copy as code in 30+ languages, New request from cURL, and history shared with the desktop app.
• Same results as the desktop app and CLI: history, cookies and settings are shared with the desktop app by default (change this with vegha.dataDirectory).

AI: Copilot, @vegha and MCP

Vegha gives language models 21 tools for your collections. They can:

• Discover: list collections, search requests, read a request.

• Change requests: create and update requests, set scripts, assertions, captures and request chains, import cURL, and create folders.

• Run: send a request or run a folder.

• Look up: read the last response, generate code, read history, and read the scripting reference.

• Copilot Chat: the tools are available in agent mode, and you can reference them with #veghaSend, #veghaRequest, and so on.

  • Every change shows its .bru diff for approval and lands as an undoable edit.
  • Every send shows its target URL before it goes out.

• @vegha participant: a request-focused assistant with /new, /test, /run, /fix, /explain and /chain.

• Other agents (Claude Code, Cursor, Claude Desktop…): run Vegha: Set Up MCP for Other Agents… to copy a ready configuration for vegha-sidecar --mcp. Those clients get the same tools; their own approval prompts apply.

Secrets never reach the model:

• Secret environment values, literal credentials in auth settings, and credential headers (Authorization, cookies, API keys) are shown as «redacted».
• Tools refuse to write that placeholder back over a real secret.
• API responses are passed to the model as untrusted data.

Getting started

New to Vegha? Follow the quick-start guide.

1. Open a folder that contains a collection (a folder with collection.bru), or run Vegha: New Collection….
2. Pick an environment from the status bar.
3. Open a request from the Vegha view in the activity bar and press Send (Ctrl+Enter).

.bru files opened from the Explorer open as text with Send and Open in Vegha Editor CodeLens links. To always use the visual editor, add "workbench.editorAssociations": { "*.bru": "vegha.requestEditor" } to your settings. Collection, folder and environment files are better edited as text.

Settings

Privacy

Vegha sends nothing anywhere except the requests you send. There is no telemetry.

Links

• Documentation: https://www.vegha.ai/vscode/
• Vegha desktop app and CLI: https://www.vegha.ai
• Issues and feature requests: https://github.com/vegha-ai/Vegha/issues

Vegha is MIT-licensed.