EXVIV Start Tracking

Public proof / not the full feed

Reviewed changes.
Source included.

This wire is a sample of P0 and P1 receipts from EXVIV’s curated public set. It publishes after an 18-hour delay. Private workspaces see the complete watchlist the same day it is reviewed.

Public receipts

305 qualifying changes.

Each row is a reviewed change with a primary source and a short before-and-after excerpt. It is proof that the machine works, not a substitute for a private watchlist.

P1 positioning

Market positioning changed: Amulet — High-performance file systems for parallel AI workloads

Amulet · Added “Amulet — High-performance file systems for parallel AI workloads”; “Instantly fork isolated, writable file systems without copying data. Mount in under 100 ms and keep work available across parallel AI workloads.”; “High-performancefile systems for parallelAI workloads”. Removed “Amulet — Google Drive for agents”; “Amulet is file storage for humans and agents—a shared place to save, find, and use files together.”.

BEFORE

Amulet — Google Drive for agents

AFTER

Amulet — High-performance file systems for parallel AI workloads

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 launch

Product release changed: Any table is a chart

Langfuse · Added “Any table is a chart”; “Toggle the Observations table into a chart over the same query. Pick a chart type, metric, aggregation, and breakdown, then add it to a dashboard.”; “Keep large observation content inspectable”. Removed “Run Experiments on Versioned Datasets”; “Fetch datasets at specific version timestamps and run experiments on historical dataset versions via UI, API, and SDKs for full reproducibility.”; “Corrected Outputs for Traces and Observations”.

BEFORE

Run Experiments on Versioned Datasets

AFTER

Any table is a chart

A new release or deprecation can change feature parity, migration risk, and the next competitive response.

P1 positioning

Braintrust shifts to active observability for production agents

Braintrust · Braintrust changed its homepage from broad AI observability to active observability for tracing, evaluating, and catching regressions in production agents.

BEFORE

Ship quality AI at scale. Braintrust is the AI observability platform for tracing production, running evals, and catching regressions before they reach users.

AFTER

Ship quality agents at scale. Braintrust is the AI observability platform for tracing production, running evals, and catching regressions before they reach users.

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 launch

Product release changed: ChatGPT Voice and multi-folder projects 26.715

OpenAI Codex · Added “ChatGPT Voice and multi-folder projects 26.715”; “Powered by GPT-Live, ChatGPT Voice lets you talk through work and coordinate tasks in Chat, Work, and Codex in the ChatGPT desktop app.”; “Start a new chat or task in voice mode, then ask ChatGPT to start, check, or steer work in other threads. On macOS, turn on Screen context to share an appshot of your frontmost win”. Removed “Reduced persistent-log churn by removing per-event WebSocket payload logging and filtering duplicated telemetry records. (#29432, #29457)”; “Full Changelog: rust-v0.141.0...rust-v0.142.0”; “#28396 [codex] Record external agent import results @charlesgong-openai”.

BEFORE

Reduced persistent-log churn by removing per-event WebSocket payload logging and filtering duplicated telemetry records. (#29432, #29457)

AFTER

ChatGPT Voice and multi-folder projects 26.715

A new release or deprecation can change feature parity, migration risk, and the next competitive response.

P1 launch

Product release changed: Cosmos Week 30 Release Notes

Augment Code · Added “Cosmos Week 30 Release Notes”; “Session forking: branch a new session from a completed turn to explore a different direction without losing the original conversation.”; “Cost Analytics: Enterprise organizations can explore billed costs with charts, detailed breakdowns, custom date ranges, filters, search, sorting, pagination, and direct session lin”. Removed “Auggie CLI 0.28.0 Release Notes”; “General-Purpose Subagent: Added a built-in general-purpose subagent option for broader multi-step research and implementation tasks.”; “Model Defaults: Added a configurable default model setting, and `/model` changes now apply only to the current session.”.

BEFORE

Auggie CLI 0.28.0 Release Notes

AFTER

Cosmos Week 30 Release Notes

A new release or deprecation can change feature parity, migration risk, and the next competitive response.

P1 positioning

Market positioning changed: Don’t Use Their Models.Use Yours.Don’t Use Their Models.Use Yours.

Understudy Labs · Added “Don’t Use Their Models.Use Yours.Don’t Use Their Models.Use Yours.”. Removed “Test open weight modelsagainst the frontier,then deploy the winner.Test open weight models against thefrontier, then deploy the winner.”.

BEFORE

Test open weight modelsagainst the frontier,then deploy the winner.Test open weight models against thefrontier, then deploy the winner.

AFTER

Don’t Use Their Models.Use Yours.Don’t Use Their Models.Use Yours.

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 enterprise

Roo Code positions Roomote as a self-hostable, single-tenant engineering agent

Roo Code · Roo Code presents Roomote as an engineering agent that ships reviewable pull requests, available hosted or self-hosted with single-tenant deployment.

BEFORE

Roomote is the AI engineering teammate you'd otherwise build from scratch or rent as a black box — source-available and self-hosted. Hand it work from Slack, an issue tracker, GitHub, or the web, bring any model, and connect the tools you already use. Get back a reviewable PR you can read end to end — solo dev or whole team.

AFTER

Roomote is an AI engineering teammate that ships reviewable PRs — hosted by us in about two minutes, or self-hosted on your own infra. Single-tenant either way, source-available, bring any model. Solo dev or whole team.

A new security, governance, or deployment capability can unlock larger accounts and reduce procurement friction.

P1 packaging

OpenRelay shifts its offer toward hosted inference and dedicated GPU compute

OpenRelay · The OpenRelay homepage now leads with hosted model inference and dedicated GPU VMs, with token-based inference and hourly GPU usage. This is a verified offer-positioning change; customer adoption impact remains unconfirmed.

BEFORE

OpenRelay — Distributed GPU cloud for production inference

AFTER

OpenRelay: Hosted inference and GPU compute

A packaging or usage-policy change can alter effective price, product eligibility, and customer comparisons.

P1 positioning

Market positioning changed: Octolane · Revenue Superintelligence

Octolane · Added “Octolane · Revenue Superintelligence”; “Octolane runs an AI agent on every account, trained on how your best revenue teams think and act. Human judgment, deployed at scale, and it's yours to keep.”; “Superintelligence is already at work”. Removed “Octolane AI”; “Self Driving AI CRM”; “Self‑driving AI CRMthat closes deals end to end”.

BEFORE

Octolane AI

AFTER

Octolane · Revenue Superintelligence

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 positioning

Market positioning changed: Reach every model, and stay up.

Keywords AI · Added “Reach every model, and stay up.”; “Ship the version that scores best.”; “Know the moment things change.”. Removed “Ship through one gateway, not a mess of moving parts.”; “Stay up when models fail”.

BEFORE

Ship through one gateway, not a mess of moving parts.

AFTER

Reach every model, and stay up.

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 positioning

Coasty shifts from medical billing to general computer-use automation

Coasty · Coasty replaced medical-billing workflow emphasis with general computer-use setup, benchmark, pricing, and native-application coverage claims.

BEFORE

Coasty is the #1 ranked computer-use AI agent (85.60% on OSWorld) and the best computer-use API of 2026 at $0.05 per /v1/predict, with 1,000+ native app integrations (Gmail, Slack, Salesforce, HubSpot, Linear, GitHub, Stripe), free sandbox keys, MCP server, and a $99 flat Unlimited plan. Cheaper than Anthropic, OpenAI, and Browserbase.

AFTER

Coasty is the #1 ranked computer-use AI agent (85.60% on OSWorld) and the best computer-use API of 2026 at $0.05 (base) per /v1/predict, with 1,000+ native app integrations (Gmail, Slack, Salesforce, HubSpot, Linear, GitHub, Stripe), free sandbox keys, MCP server, and a $99 flat Unlimited plan. Cheaper than Anthropic, OpenAI, and Browserbase.

An ecosystem expansion can unlock distribution, increase switching costs, or close an adoption gap.

P1 positioning

Market positioning changed: Almanac | The context layer for engineering teams

Almanac · Added “Almanac | The context layer for engineering teams”; “A single source of truth for your codebase, so every developer and coding agent stays on the same page.”; “The context layer forengineering teams”. Removed “Almanac — A living wiki for your codebase”; “Every coding session leaves context behind. Almanac saves it in a wiki.”; “A living wiki for yourcodebase”.

BEFORE

Almanac — A living wiki for your codebase

AFTER

Almanac | The context layer for engineering teams

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P0 packaging

OpenAI adds separate Voice and Codex usage meters

OpenAI Codex · For ChatGPT Business and Enterprise customers using credits or pay-as-you-go billing, Desktop Voice is expected to cost about 6 credits per connected minute. Codex tasks started by Voice also consume the shared Codex credit pool separately at token-based rates.

BEFORE

Managing billing and seats in ChatGPT BusinessLearn how standard ChatGPT seats, Codex seats, credits, and billing cadence work in ChatGPT Business.

AFTER

ChatGPT Voice in Desktop

A packaging or usage-policy change can alter effective price, product eligibility, and customer comparisons.

P1 reliability

Service reliability changed: We’re fully operational

Bubble · Added “We’re fully operational”. Removed “We’re currently experiencing issues”; “We've seen a spike of unidentified traffic which has now been subsided. We are making efforts to prevent this in the future.”.

BEFORE

We’re currently experiencing issues

AFTER

We’re fully operational

A material reliability change can affect buyer confidence, renewal risk, and competitive displacement opportunities.

P1 integration

Zed adds Git grouping, multi-select finders, and CSV filtering

Zed · Zed added staged and unstaged Git grouping, multi-select in file and text finders, CSV row filtering, and improved branch filters.

BEFORE

This week's release includes automatic agent context compaction that you can also trigger manually with the new /compact command, cleaner and more legible Markdown preview styling, and support for running custom git commands on git graph branches and tags.

AFTER

This week's release includes Staged and Unstaged grouping in the Git Panel, multi-select in the File Finder and Text Finder, and row filtering in CSV previews.

An ecosystem expansion can unlock distribution, increase switching costs, or close an adoption gap.

P1 positioning

Market positioning changed: Supernova.io — Design system and knowledge management platform

Supernova.io · Added “Supernova.io — Design system and knowledge management platform”; “Supernova.io is the agentic design system platform that turns your tokens, components, documentation, and code patterns into connected intelligence — so your team and their AI agen”. Removed “Supernova.io — AI-powered platform for product teams”; “Supernova.io is a collaborative platform that empowers product teams to create and write content with AI assistance while staying connected to their company's design system.”.

BEFORE

Supernova.io — AI-powered platform for product teams

AFTER

Supernova.io — Design system and knowledge management platform

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 positioning

Quantstruct pivots from document automation to hotel housekeeping robots

Quantstruct · Quantstruct replaced AI documentation-automation positioning with hotel housekeeping robots for linen handling, bed turnover, and bathroom cleaning.

BEFORE

AI-powered documentation automation

AFTER

Robots for linen handling, bed turnover, and bathroom cleaning.

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 positioning

Popl broadens from event lead capture to an in-person GTM platform

Popl · Popl broadened its headline from event lead capture to an AI GTM platform for in-person events and raised its displayed adoption claim from 2.5M to 4M+ professionals.

BEFORE

Your AI GTM platformfor event lead capture

AFTER

Your AI GTM platformfor in-person events

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 positioning

Market positioning changed: Perspectives

Perspectives Health · Added “Perspectives”; “Perspectives runs your behavioral health program's back office: nightly chart audits, UR generation, clinical documentation, and payer audit defense.”; “Perspectives helps run your behavioral health program to its full potential”. Removed “Perspectives Health — AI for behavioral health”; “Perspectives runs your behavioral health clinic to its full potential”; “Nightly Audits”.

BEFORE

Perspectives Health — AI for behavioral health

AFTER

Perspectives

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 positioning

Million shifts from code verification to developer tools that train models

Million · Million replaced a code-verification proposition with open-source developer tools and environments intended to help models create a better web.

BEFORE

Million is solving the hardest problem in AI coding: proving the code actually works.

AFTER

Million builds widely adopted open-source developer tools and environments that teach models to create a better web.

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 positioning

GhostEye expands from social-engineering tests to end-to-end attack validation

GhostEye · GhostEye broadened from authorized social-engineering exercises to attacks that continue into customer systems and validate detections and endpoint defenses.

BEFORE

GhostEye runs authorized, AI-driven social-engineering attacks on your people and hands security teams the exact gaps to fix — the modern alternative to security-awareness training.

AFTER

GhostEye runs authorized attacks on your people, follows them into your systems, and validates whether your detections and EDR catch what gets through.

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

P1 positioning

Market positioning changed: Anakin.io — One API. The whole web.

Anakin · Added “Anakin.io — One API. The whole web.”; “Turn any website into clean, agent-ready data. Scrape, crawl, search, and run pre-built actions across 900+ sites. One key, structured JSON, 99.9% uptime.”. Removed “Anakin.io - Lightning Fast Web Scraping API”; “Scrape any page with zero blocks and lightning-fast performance. 30x faster with caching, 99.9% uptime, enterprise-ready web scraping API.”.

BEFORE

Anakin.io - Lightning Fast Web Scraping API

AFTER

Anakin.io — One API. The whole web.

A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.

Need the feed for the companies you actually sell against?

Start with Scout for one private competitor. Pulse and Radar cover the rest of the watchlist, same day, with exports.

Start Tracking