Who it is for
Engineering and support teams already running agents on bug reports who want those agents inspecting real sessions rather than summaries.
Overview
What Subtext does
Fullstory's Subtext exposes recorded production sessions as something an agent can query rather than a video a person watches. An agent is handed a session link, often from a support thread, a Sentry issue, a Linear ticket, or a pull request, then works through it with dedicated tools: jumping to a specific moment, comparing DOM state across it, pulling screenshots and accessibility trees, and reading windowed console and network context.
The tool surface is shaped to keep token use down compared with pointing a general agent at raw replay data. Privacy handling covers field masking and exclusion controls plus DLP scanning exposed over MCP. Capture runs without sampling, and agents review the subset that matters.
It works with Codex or other MCP-compatible agents, installs from code, and is priced by usage with a free tier.
What stands out
Session data arrives as agent-callable tools rather than a timeline to scrub, so the investigation happens without anyone watching the recording.
Availability
Where Subtext is listed
Fullstory ships this on both marketplaces: Subtext is listed in the Claude connectors directory and in the ChatGPT apps directory, under the same publisher domain.
Only 392 of the 3,150 products in this directory are listed in both marketplaces. One MCP server can serve both surfaces; what differs is each platform's listing and review process.
ChatGPT listing
In the ChatGPT apps directory
- Developer
- Fullstory
- Filed under
- Data & Analytics
- Listed since
- July 2026
- Authentication
- Authenticates when installed
- Website
- subtext.fullstory.com
- Policies published
- Privacy policy · Terms of service
- Official listing
- Open in ChatGPT
- Snapshot
- August 2026
Named skills
Subtext declares 6 named skills, which puts it in the 6.7% of ChatGPT-listed apps that do.
- subtext-privacy
- subtext-review
- subtext-session
- subtext-shared
- subtext-telemetry
- subtext-using-subtext
Example prompts from the listing
- Review this user session and provide an analysis of what happened
- Review recent user sessions and provide ideas for UX improvements
- Find the root cause for this customer-facing bug
Claude listing
In the Claude connectors directory
Subtext publishes 15 tools to Claude. That is a mid-sized surface: 59.4% of the 1,253 Claude-listed connectors expose fewer.
The median Claude-listed connector filed under Developer Tools publishes 13, so this one runs 2 tools above its bucket's median and ranks 43rd of 95 by tool count.
It falls in the 11-20 tools band, which holds 22.5% of the Claude directory.
Subtext publishes a remote MCP endpoint. Claude connects to it over HTTP; there is nothing to install locally.
Every connector in the Claude directory is remote, all 1,253 of them, so transport is not a differentiator there.
- Publisher
- Fullstory
- Verification
- Verified partner (37.8% of the Claude directory)
- Filed under
- Developer Tools
- Endpoint
-
https://api.fullstory.com/mcp/subtext - Documentation
- Publisher setup docs
- Published tools
- 15 · 11-20 tools band
- Depth rank in Developer Tools
- 43 of 95
- Snapshot
- August 2026
Subtext carries partner verification in the Claude directory, which 474 of 1,253 listings do (37.8%). That is a review step past self-publishing, not an Anthropic-built server.
In the publisher's words
How Fullstory describes it
**Subtext is session replay built for agents instead of humans.** Connect it and your Claude agents can open real production sessions and investigate them directly—no scrubbing a timeline, no video to watch. Hand your agent a Subtext link from wherever it already looks—a Sentry issue, a Linear ticket, a support thread, a PR—and it jumps to the right moment, inspects screenshots and accessibility trees, diffs DOM state, and reads console and network activity to find the root cause. Then it files the fix or hands you the finding. You review the answer, not the replay. What your agent gets: * **Session tools it calls directly**: open a session, jump to a moment, inspect and compare state—purpose-built so agents traverse dense session data efficiently, with fewer tokens than pointing a general agent at raw replay. * **Best-in-class capture from Fullstory**: Fullcapture behavior, pixel-perfect screenshots on demand, DOM diffs, and windowed console and network context. * **Privacy built for production**: precision tools for agents to mask or exclude sensitive fields, plus agent-friendly DLP scanning tools over MCP. * **Cost-effective**: Capture every session without sampling and never miss a critical user experience or bug report. Agents review the ones that matter most. Works with Claude Code or any MCP-compatible agent. Code-first install, self-serve signup, generous free tier, usage-based pricing. Bring your own agent harness—Subtext is the session context it was missing. Session replay used to be something humans watched. Now it's evidence your agents act on.
Quoted from the connector's own Claude directory listing as captured in August 2026. Node8 did not write it and does not vouch for it.
Nearby
Other Developer Tools connectors
Botpress
Agent prototyping on the Botpress platform
ChatGPT
Mnemom
Trust ratings and identity for software agents
Claude · ChatGPT
Postman
API collections, environments, and workspaces from Postman
Claude
Resolve AI
Production investigations and root-cause work inside coding agents
Claude · ChatGPT
Retool
Retool workspace administration and internal app management
Claude
Sourcegraph
Semantic code search and symbol lookup across many repositories
Claude
This is an independent catalogue entry compiled by Node8 from Subtext's public marketplace listings as they stood in August 2026. Listing facts come from the marketplaces; the comparisons against the rest of the directory and any live-endpoint measurements are Node8's. Node8 is not affiliated with Fullstory, and inclusion here is not an endorsement.
Want your company reachable inside ChatGPT and Claude?
Node8 builds MCP servers end to end: deciding which tools are worth exposing, the authentication and permission model, the review submissions, and the onboarding that gets them used. One server serves ChatGPT, Claude, and Microsoft Copilot.