# VS Code: see who AI answers name instead of you, in Copilot Chat

> For the complete documentation index, see [llms.txt](https://attensira.com/llms.txt).
> Every page on this site is also served as Markdown: append `.md` to any URL, or send `Accept: text/markdown`.

> Add Attensira's MCP server to VS Code with a three-line mcp.json, sign in from the Start link, and use it from Copilot Chat in agent mode.

- Canonical page: https://attensira.com/integrations/vs-code
- Family: integrations
- Last updated: 2026-10-04
- Author: Karl-Gustav Kallasmaa (https://x.com/kallasmaa)
- Claims as data: https://attensira.com/claims.json?slug=vs-code

**In short:** Add a three-line .vscode/mcp.json pointing at https://mcp.attensira.com/mcp, press Start above the entry and sign in in the browser. VS Code keeps the token in its own account storage, not the file. Then open Copilot Chat in agent mode and ask which rivals AI answers name instead of you, on which questions, and what the crew has waiting.

## What it showed on attensira.com (read 2026-10-04)

Asked in VS Code · Copilot Chat: "Head to head, how often is each rival named in an answer that leaves us out?"

Call: `get_analytics area=competitor days=30`

Of 4,518 answers in 30 days, Semrush was named without us in 1,601. We were named without Semrush in 177.

| Row | Value | Note |
| --- | --- | --- |
| Semrush named, we were not | 1,601 | we without them 177 · both 20 |
| Profound named, we were not | 1,333 | we without them 113 · both 84 |
| Ahrefs named, we were not | 1,155 | we without them 181 · both 16 |
| Peec AI named, we were not | 972 | we without them 184 · both 13 |

_Every answer counted once per rival, out of 4,518 read._

## What can Copilot Chat do with it?

- Read, per rival, how often answers name them and not you, how often they name you and not them, and how often both, out of every answer read.
- Open the prompts where a rival wins and read what the model actually said about them, before deciding what to change.
- Keep a project-level mcp.json in the repository so every collaborator gets the server and signs in as themselves.
- Approve the crew's waiting work or hand it a fix from Copilot Chat, marked "via VS Code" in the app.

## How do I connect VS Code?

1. **Write the server entry** VS Code reads .mcp.json at the workspace root or .vscode/mcp.json. The entry names the server, sets the http type and gives the URL. With OAuth the file carries nothing secret, so it is safe beside the code.
2. **Press Start and sign in** Use the Start link above the entry. VS Code opens a browser for sign-in and stores the token in its own account storage, not in the file. Approve, and the connection reaches every workspace your organisation owns.
3. **Open Copilot Chat in agent mode** Tools only appear to an agent. Open Copilot Chat in agent mode and check the tools picker. If Attensira is not listed, the server has not started, which is different from a failed sign-in.
4. **Check it with get_account** Ask Copilot to call get_account. It lists your workspaces, the credit balance, the scopes and the platforms really measured. If nothing answers, https://mcp.attensira.com/healthz tells you whether the server is reachable.

## What are the limits in VS Code?

- Head-to-head counts cover only the rivals you track. A competitor you have not added reads as absent, which is not the same as not appearing.
- A signed-in connection carries read and write: Copilot can add prompts or spend credits if asked. Use the tools picker to limit it, or a read-only API key.
- A rival named more often than you is not proof they are gaining. Gaining is a change over time; read get_analytics area=competitor over two windows and trust only deltas marked real.

### What did the head-to-head show on our own workspace?

Every tracked rival was named without us far more often than we were named without them. That is not a reason to despair, it is a work queue: open the prompts behind those counts and the crew can see which pages and lists the models lean on when they name a rival and not us.

### How do you read a head-to-head honestly?

Three cautions keep these numbers useful. They only cover rivals you track: one you have not added reads as absent. They are counts over a window, not a trend: "named more often than us" is not "gaining", and gaining is a change read over two windows, trusting only deltas marked `real: true`. And answers name several brands at once, so the counts for different rivals overlap and never add up to the total.

## FAQ

### Where does VS Code look for the server entry?

At .mcp.json in the workspace root, or .vscode/mcp.json. With OAuth the whole entry is three lines and carries nothing secret.

### Where is the token stored?

In VS Code's own account storage, not in the file. That is why the file is safe to commit.

### I do not see Attensira in Copilot Chat.

Tools only appear in agent mode. Open Copilot Chat in agent mode and check the tools picker. If it is still missing, the server has not started; use the Start link above the entry.

### Can I use a key in VS Code?

Yes, as an Authorization header, but keep it out of any repo-visible file. On a machine with a browser, signing in is the better path.

### Does a head-to-head read cost credits?

No. Every read is free however often you ask. Only tools that start agent work, such as ask, run_automation and add_prompts, spend credits.

## Sources

Each entry below is one claim on this page. `retrieved` is the date the exporter or a human last saw the source say this, and nothing more: it is not a publication date and not a promise of re-verification. `cite` is the anchor on the canonical page where this claim is rendered.

### vscode-token-in-account-storage

Attensira's connect documentation states that VS Code opens a browser for sign-in and stores the resulting token in its own account storage rather than in the configuration file.

- source: https://docs.attensira.com/mcp/connect (docs.attensira.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/integrations/vs-code#claim-vscode-token-in-account-storage

> VS Code opens a browser for sign-in and stores the resulting token in its own account storage, not in the file.
### vscode-config-locations

Attensira's connect documentation states that VS Code reads mcp.json at the workspace root or at .vscode/mcp.json.

- source: https://docs.attensira.com/mcp/connect (docs.attensira.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/integrations/vs-code#claim-vscode-config-locations

> VS Code reads `.mcp.json` at the workspace root or `.vscode/mcp.json`.
### mcp-healthz-unauthenticated

Attensira's MCP overview states that the server's health path is unauthenticated and answers without a key, making it a useful reachability check.

- source: https://docs.attensira.com/mcp/overview (docs.attensira.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/integrations/vs-code#claim-mcp-healthz-unauthenticated

> is unauthenticated and answers without a key, which makes it a useful reachability check when a client is silent about why it failed