Skip to content

The cross-harness, self-improving memory layer for AI agents. Join our community: https://discord.com/invite/zdNChS2fBu

docs(browser-extension): state that reasoning is not available on claude.ai or ChatGPT (#1013)

Shukan Shah69df98582012
diff --git a/docs/runtimes/browser-extension.mdx b/docs/runtimes/browser-extension.mdx
index 41e787f4..76d3fc60 100644
--- a/docs/runtimes/browser-extension.mdx
+++ b/docs/runtimes/browser-extension.mdx
@@ -207,6 +207,7 @@ Coverage that is the same for both sites:
| Prompt telemetry | Supported, with full prompt text under `full` and `redacted` retention |
| Turn identity | Supported. Each chat request's turn id is sent as `beacon.prompt.id` and lands in `prompt.id` |
| Assistant response telemetry | Supported, including the response text and model |
+| Reasoning telemetry | Not available. Neither site sends the model's reasoning to the page, only short summaries of it, and those summaries are not recorded |
| Command and file telemetry | Not applicable. Browser chat performs no local commands or file edits |
| Approval telemetry | Not available. Neither site exposes approval decisions to a page script |
| Session lifecycle | Partial. Turns are correlated by conversation id; there is no session start or stop signal |
diff --git a/docs/runtimes/chatgpt-web.mdx b/docs/runtimes/chatgpt-web.mdx
index 91b2060d..17cf71a1 100644
--- a/docs/runtimes/chatgpt-web.mdx
+++ b/docs/runtimes/chatgpt-web.mdx
@@ -54,6 +54,7 @@ Turns from one conversation share a session id, so a conversation reads as a ses
|------|---------|
| Prompt telemetry | Supported, with full prompt text under `full` and `redacted` retention |
| Assistant response telemetry | Supported, including the response text, the requested model, and the model that answered |
+| Reasoning telemetry | Not available. ChatGPT sends the page only a short summary of its reasoning, and Beacon does not record it |
| Tool telemetry | Partial. Web searches are captured with their results, and with their queries when the site reports them, which it does not do for every search. Other built-in tools are recorded by name, with any result text they stream |
| Token usage | Not available. ChatGPT web does not report token counts in its stream |
| Cost | Not available. ChatGPT web does not report per-turn cost to the page |
diff --git a/docs/runtimes/claude-web.mdx b/docs/runtimes/claude-web.mdx
index c80c8aa2..ad1beb72 100644
--- a/docs/runtimes/claude-web.mdx
+++ b/docs/runtimes/claude-web.mdx
@@ -54,7 +54,7 @@ Turns from one conversation share a session id, so a conversation reads as a ses
| Assistant response telemetry | Supported, including the response text and the model reported by the stream |
| Tool telemetry | Supported. Tool call id, name, and arguments are captured from `tool_use` blocks |
| Tool results | Supported for tools claude.ai runs on its side, such as web search and weather. The stream sends each result as a `tool_result` block, recorded in `gen_ai.tool.call.result` |
-| Reasoning | Not captured. claude.ai streams short thinking summaries, which the adapter skips |
+| Reasoning telemetry | Not available. claude.ai sends the page only a one-line summary of each thinking block, and Beacon does not record it |
| Token usage | Parsed when the stream carries it, which recorded traffic so far does not. See [Limits](#limits) |
| Cost | Not available. claude.ai does not report per-turn cost to the page |
| Command and file telemetry | Not applicable. Browser chat performs no local commands or file edits |