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.
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.
Speechmatics · Speechmatics now leads with low-latency speech-to-text for multilingual, multi-speaker conversations and emphasizes voice-agent builders, real-world audio accuracy and flexible cloud, on-premise and on-device deployment.
BEFORE
Speechmatics offer the most accurate AI speech technology for enterprise - with AI transcription, real-time translation and text-to-speech components. Try our Speech API today!
→
AFTER
A speech-to-text API for real conversations, not clean demos. Build Voice AI that understands accents, multiple speakers, and multilingual speech.
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Unsloth · Unsloth Desktop runs and trains models locally on Mac, Windows and Linux, adds image and video generation, connects Claude Code and Codex to local models, and includes private web search and sandboxed code execution.
BEFORE
Unsloth - Train and Run Models Locally
→
AFTER
Unsloth - Run and Train Models Locally
An ecosystem expansion can add distribution, increase switching costs, or close an adoption gap.
Axiom.ai · Axiom.ai exposed its browser-automation steps through a Node library, MCP server and existing Puppeteer or Playwright sessions, including a one-line scraper with automatic pagination and retries.
BEFORE
Creating a Reddit Notification Bot
→
AFTER
Web scraping in one line of code (or zero)
An ecosystem expansion can add distribution, increase switching costs, or close an adoption gap.
KugelAudio · KugelAudio now markets Kugel 3 for real-time voice applications in 26 languages, down from its prior 40-plus-language claim, and adds direct EU hosting plus customer-operated and on-premise deployment options.
BEFORE
KugelAudio provides production-ready text-to-speech for voice AI agents. Its voices are developed and hosted in Europe, support 40+ languages, and are designed for GDPR-compliant deployments.
→
AFTER
KugelAudio provides text-to-speech for real-time voice applications. Kugel 3 supports 26 languages through a hosted API, with a direct EU endpoint and customer-operated deployment options.
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Prox · Prox now positions its product specialist around selection, configuration, competitive guidance, quoting and calculations across websites, dealer portals, QR codes, messaging and MCP, with a stated two-week launch process.
BEFORE
Don't make customersdig through manuals.
→
AFTER
Doka
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Replicas · Replicas now starts coding-agent work from Linear, Slack, GitHub, GitLab, its dashboard, schedules, webhooks or API calls and supports Claude Code, Codex, Cursor and Opencode in isolated Linux machines.
BEFORE
Assign the task
→
AFTER
Assign from anywhere
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Levocred AI · Levocred now targets credit funds, fintechs and banks with an AI employee for underwriting, portfolio monitoring, compliance and reporting, and says roughly $1 billion in receivables runs through the platform.
BEFORE
Levocred | OS for Structured Credit
→
AFTER
Levocred | AI for credit funds, fintechs and banks
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Hindsight · Hindsight now builds a verified memory from calls, emails, CRM records and warehouse data, then applies comparable deal experience across teams, agents and workflows.
BEFORE
Hindsight | Every Deal, Scored Against Your Playbook
→
AFTER
Hindsight | Turn Every Deal Into an Advantage
An ecosystem expansion can add distribution, increase switching costs, or close an adoption gap.
OpenAI Codex · Daybreak Blue provides approved defenders access to general-purpose models for defensive security work, while Daybreak Red requires separate approval for purpose-trained models used in authorized exploit validation and red teaming.
BEFORE
#34232 Remeasure dynamic cells in the transcript overlay @copyberry
→
AFTER
Introducing Daybreak Blue and Daybreak Red
A new release or deprecation can change feature parity, migration risk, and the next competitive response.
Intryc · Intryc now connects quality assurance, root-cause insights, coaching, training and automation so customer-interaction findings can drive measurable improvement.
BEFORE
AI QA and Training Platform for Human and AI CX Agents
→
AFTER
Intryc: Close the Loop on Customer Experience
An ecosystem expansion can add distribution, increase switching costs, or close an adoption gap.
Lindy · Lindy is now positioned as an AI employee that connects to company tools, uses organization-wide context and performs work for a whole team rather than primarily managing one user's inbox and calendar.
BEFORE
Lindy – The Ultimate AI Executive Assistant
→
AFTER
Lindy – The AI Teammate That Lets You Do More
An ecosystem expansion can add distribution, increase switching costs, or close an adoption gap.
Pollen · Pollen now offers named AI employees that learn and execute workflows through email and chat for advisors, accountants, investors, insurers, lenders and family offices.
BEFORE
Pollen · Automation for document-heavy client workflows
→
AFTER
Pollen · Your AI employee for financial workflows
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Sim · Sim v0.7.64 adds Snowflake credential authentication, object pickers, nine operations and PAT support, plus a Mintlify integration and PI plan mode.
BEFORE
v0.7.55
→
AFTER
v0.7.64
An ecosystem expansion can add distribution, increase switching costs, or close an adoption gap.
Clarm · Clarm now targets operations teams building production coworkers from IT- and compliance-approved sources, actions, memory and tools, with auditable execution receipts.
BEFORE
Clarm - Governed AI agents for regulated teams
→
AFTER
Clarm Atlas - AI coworkers for operational work
A new security, governance, or deployment capability can open larger accounts and reduce procurement friction.
Orange Slice · Orange Slice now highlights ready-made workflows for hiring research, competitive intelligence, social-media research and ecommerce research alongside lead prospecting and data enrichment.
BEFORE
Refer customers. Earn 20% for 12 months.
→
AFTER
What teams build with Orange Slice
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Agnost AI · Agnost AI now analyzes production chat and voice conversations for silent failures, user frustration, feature requests, policy violations, and reviewed improvement opportunities.
BEFORE
Agnost AI | Make your agent better
→
AFTER
Agnost AI | Product analytics for conversational agents
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
MarkIt · MarkIt now connects product evidence, HS and HTS classification, trade documents, reporting, expert review, and auditable compliance decisions in one workspace.
BEFORE
MarkIt | AI HS & HTS Code Classification for Trade Compliance
→
AFTER
Trade Compliance Software for Global Teams | MarkIt
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Warp · Warp now positions its terminal, Warp Agent, and Oz as an open platform spanning local agent workflows, cloud software factories, model choice, governance, and usage controls.
BEFORE
Warp — The Agentic Development Environment
→
AFTER
Warp — The open platform for automating development
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Trigger.dev · Trigger.dev now provisions and revokes organization membership automatically when administrators assign or unassign users through their identity provider.
BEFORE
Trigger.dev v4.4.6
→
AFTER
SCIM directory sync for SSO
A new security, governance, or deployment capability can open larger accounts and reduce procurement friction.
BLACKBOX AI · BLACKBOX AI now says Enterprise Inference and its 300-plus-model Router use end-to-end encryption, replacing prior copy that described encryption in transit.
BEFORE
Private,secureinferenceatrecordspeed.
→
AFTER
Endtoendencryptedinferenceforeverymodel.
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Magnitude · Magnitude now positions its local-model runtime as a general desktop agent for sensitive data, private notes, code, logs, files, documents, browser work, and automation.
BEFORE
Code with local models, zero setup | Magnitude
→
AFTER
Your actually local agent | Magnitude
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Bolt.new · Bolt.new says it deployed a fix for the intermittent project-publishing issue and is monitoring the result while a separate upstream-provider outage remains active.
BEFORE
We've identified and are working on a fix for the issue where publishing can hang on "Publishing your site, this will just take a moment…" and never finish. This is intermittentand affects individual projects, so your other projects may publish normally.In the meantime, if a publish gets stuck, refresh the page and click Publish again straight away. That should publish your website in most cases.
→
AFTER
Upstream Provider has reported an outage
A material reliability change can affect buyer confidence, renewal risk, and competitive displacement opportunities.
Ellipsis (YC W24) · Ellipsis now supports sandboxed agent fleets with scoped credentials, hard budgets, and audit trails, plus code-review pipelines using customer-selected prompts, models, reviewers, and filters.
BEFORE
Use cases
→
AFTER
A fleet of cloud agents, managed like infrastructure
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.
Bolt.new · Bolt.new says the publishing hang affects individual projects intermittently and recommends refreshing the page and publishing again while it works on a fix.
BEFORE
We are investigating an issue where some users may experience publishing hanging/getting stuck with "Publishing your site, this will just take a moment..."
→
AFTER
We've identified and are working on a fix for the issue where publishing can hang on "Publishing your site, this will just take a moment…" and never finish. This is intermittentand affects individual projects, so your other projects may publish normally.In the meantime, if a publish gets stuck, refresh the page and click Publish again straight away. That should publish your website in most cases.
A material reliability change can affect buyer confidence, renewal risk, and competitive displacement opportunities.
Inventive AI · Inventive AI now positions specialized agents across go/no-go evaluation, response drafting, compliance review, content governance, competitive positioning, and final export.
BEFORE
Simplify RFP Workflows With AI RFP Software
→
AFTER
Autonomous AI Agents That Run Your RFPs End to End
A messaging shift can reveal a new buyer, category claim, use case, or go-to-market emphasis.