Verified AI Launch Intelligence
See what launched, what changed, and what it costs.
Track AI product launches, model updates, pricing changes and tested workflows. Each record separates official sources, company claims, Kingy testing and third-party evidence so you can see what is known and what remains unverified.
Daily radar
Today's AI Launches
A Radar edition publishes only when the day has enough source-ready signal; the tracker continues to show the newest verified records.
Daily AI Launch Radar: July 24, 2026
Daily Kingy AI Launch Radar for 2026-07-24: source-checked AI launches, product updates, pricing notes, use cases, and risks worth watching.
No source-ready records are tagged for today yet, so this section is showing the latest public launch records available in the tracker.
Microsoft Project Perception
Summary: Microsoft announced Project Perception, an agentic security system that coordinates specialized red, blue, and green agents inside Microsoft Defender.
Why it matters: The launch brings attack simulation, defense testing, and threat-scenario planning into one Microsoft Defender workflow. Its practical effectiveness and cost remain unproven until security…
- Pricing
- Microsoft says Project Perception uses consumption-based, pay-as-you-go pricing measured in Security Compute Units (SCUs). Different agents consume SCUs at different rates; Microsoft did not publish numeric SCU rates in the checked launch sources.
- Verification
- Verified; 3 sources
Anthropic deprecates non-default sampling controls on newer Claude models
Summary: Anthropic deprecated non-default temperature, top_p and top_k values on Claude Opus 4.7 and later models. Requests that set them now return HTTP 400; Anthropic says to…
Why it matters: This is a request-compatibility break for teams that carry sampling overrides forward between Claude versions. Remove non-default temperature, top_p and top_k values before moving…
- Pricing
- The parameter deprecation has no separate charge. Claude API usage is billed per token at model-specific rates; Anthropic currently lists Claude Opus 4.7 at $5 per million base input tokens and $25 per million output tokens. New users receive a small amount of free API credits.
- Verification
- Verified; 2 sources
kausable raises €12 million for adaptive causal AI
Summary: Heidelberg-based kausable raised a verified €12 million seed round to develop causal AI models intended to adapt to new contexts with minimal data.
Why it matters: kausable seed round addresses a concrete need: The funding backs a technically different approach to adaptive AI, but claims about eliminating retraining and handling…
- Pricing
- No public pricing was verified in this run.
- Verification
- Verified; 1 source
Google opens Gemini Startup Forum Winter 2026 applications
Summary: Google opened applications for a two-day Gemini Startup Forum for seed-to-Series-A companies building with AI.
Why it matters: Gemini Startup Forum Winter 2026 addresses a concrete need: The program is a practical access point for early-stage AI startups that need technical guidance…
- Pricing
- No application fee or program price was stated in the official announcement.
- Verification
- Verified; 2 sources
Health in ChatGPT connects U.S. health and wellness data
Summary: OpenAI launched a U.S. rollout that lets eligible ChatGPT users connect health records and supported wellness data for more contextual health conversations.
Why it matters: Health in ChatGPT addresses a concrete need: Connected records make health conversations more personally relevant, but the launch also raises practical questions about rollout…
- Pricing
- Health in ChatGPT is included within ChatGPT access rather than listed as a separate paid product. OpenAI says GPT-5.5 Instant is available on the Free plan and GPT-5.6 Sol is available to paid users; current ChatGPT plan pricing should be checked separately.
- Verification
- Verified; 3 sources
GitHub Issues adds agent automation controls
Summary: GitHub added public-preview approval, confidence, and rationale controls for agent-driven changes to issue labels, fields, types, status, and assignees.
Why it matters: Agent automation controls in GitHub Issues addresses a concrete need: The controls make issue automation more observable and reviewable, but GitHub explicitly says approvals…
- Pricing
- GitHub did not publish a separate price for the public-preview controls. Availability depends on eligible GitHub and Copilot plans and organization settings.
- Verification
- Verified; 4 sources
Kingy Launch Brief
Put the week’s verified AI launches in your inbox.
One source-checked edition every Friday, with a clear try, watch or skip verdict. After subscribing, check your inbox and confirm your address.
Free · Fridays · Double opt-in · Unsubscribe anytime
Weekly awards
This Week's Strongest AI Launches
The weekly awards edition is editorial, source-aware, and designed to separate useful launch signal from announcement noise.
Kingy AI Launches of the Week: June 15-19, 2026
Kingy AI Launches of the Week for June 15-19, 2026: editorial awards for GitHub Agent Finder, Microsoft Copilot Cowork, GitHub Copilot App, Gemini Omni, MolmoMotion, ChatGPT Scheduled Tasks, and the…
Category navigation
Find AI launches by market, workflow, and source signal
Launch tracker
Browse the AI Launch Tracker
Search source-backed launch records, then use common filters first and advanced filters only when you need a narrower view.
Anthropic deprecates non-default sampling controls on newer Claude models
Anthropic deprecated non-default temperature, top_p and top_k values on Claude Opus 4.7 and later models. Requests that set them now return HTTP 400; Anthropic says to omit the…
This is a request-compatibility break for teams that carry sampling overrides forward between Claude versions. Remove non-default temperature, top_p and top_k values before moving…
Google lists October 16 as Gemini 2.5 Pro’s earliest possible shutdown date
Google lists October 16, 2026 as the earliest possible shutdown date for Gemini 2.5 Pro and names gemini-3.1-pro-preview as the recommended replacement. Google says users will receive advance…
Treat October 16 as a migration deadline to plan around, not a guaranteed retirement day. Google labels it the earliest possible shutdown date for…
Gemini changes sampling and turn-validation rules for new models
Google deprecated temperature, top_p and top_k for Gemini 3.6 Flash, Gemini 3.5 Flash-Lite and later Gemini models. The named models ignore those parameters; future generations will return HTTP…
Remove these legacy controls before migrating to Gemini 3.6 Flash, Gemini 3.5 Flash-Lite or later releases. Sampling parameters are ignored on the current named…
LeRobot v0.6.0
Hugging Face released LeRobot v0.6.0 with world-model policies, a reward-model API, six new simulation benchmark integrations, a deployment rollout CLI, DAgger-style human correction, richer dataset tooling, FSDP and…
LeRobot v0.6.0 connects evaluation, deployment, intervention data and retraining more coherently than earlier releases, and its source material is unusually detailed. Breadth is also…
Elastic Training with MaxText
Google published an end-to-end elastic-training workflow for MaxText, Pathways, GKE, and Cloud TPUs on July 6, 2026.
- Kingy
- 7.1 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
Google published an end-to-end elastic-training workflow for MaxText across Pathways, GKE, and Cloud TPUs, turning a mid-training slice failure into a recoverable event (developers.googleblog.com).…
Hugging Face Kernels
Hugging Face published a major Kernels redesign on July 6, 2026, including a first-class Hub repository type for compute kernels, stricter publisher controls, redesigned command-line tools, and broader…
- Kingy
- 7.7 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
Hugging Face shipped a major Kernels redesign — a first-class Hub repository type for compute kernels, stricter publisher controls, redesigned command-line tools, and broader…
Genkit Agents API
Google introduced the Genkit Agents API in preview for TypeScript and Go, with a shared chat interface, streaming, server- or client-managed state, snapshots, human interrupts, detached tasks and…
- Kingy
- 7.6 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
Genkit Agents API removes repeated full-stack agent plumbing while leaving teams in control of runtime and state ownership. The preview can introduce breaking changes…
Kotoba Technologies $10M Seed Extension
Kotoba Technologies announced an additional $10 million seed investment led by Kindred Ventures, with Salesforce Ventures and Sony Innovation Fund participating, bringing disclosed total funding to $23 million…
- Kingy
- 5.7 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
Kotoba’s additional $10 million is a meaningful capacity signal for a technically specific East Asian voice strategy, not proof that its models outperform alternatives.…
Jan v0.8.3 release
Jan released v0.8.3 with fixes across MCP browser-port handling, backend installs, media handling, and chat rendering.
- Kingy
- 7.4 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
Jan's v0.8.3 release on GitHub fixes MCP browser-port handling, backend installs, media handling, and chat rendering in its open-source, local-first desktop assistant (github.com/janhq/jan). It…
Latitude V2 Agent Monitoring
Latitude launched V2 for AI-agent monitoring, combining production traces, session search, behavior-pattern signals, recurring-failure discovery, alerts and a coding-agent handoff supplied with issue context.
- Kingy
- 7.6 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
Latitude V2 creates a useful closed loop from production traces to recurring-failure signals and a proposed repair. Kingy did not connect telemetry, measure cluster…
Cloudflare Temporary Accounts for AI Agents
Cloudflare added 60-minute Temporary Accounts to Wrangler so an unauthenticated coding agent can deploy and revise supported Workers resources, return a preview and claim URL, and let a…
- Kingy
- 7.6 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
Temporary Accounts remove real signup friction and delete unclaimed previews after 60 minutes. The claim URL and temporary token are sensitive values, so production…
Gemini Enterprise Workflow Agents
Google made Gemini Enterprise Workflow Agents generally available with an allowlist, enabling authorized users to create, import, update and run triggered sequences that mix AI automation, connected actions…
- Kingy
- 6.6 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
Gemini Enterprise Workflow Agents offer a more governable structure than an improvised multi-step chat, but GA does not mean open access: the Cloud project…
Ollama v0.30.10 release
Ollama released v0.30.10 with Apple Silicon MLX support for Command A and North family models, plus llama.cpp and build-artifact updates.
- Kingy
- 7.7 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
Ollama's v0.30.10 release adds Apple Silicon MLX support for the Command A and North model families, plus llama.cpp and build-artifact updates, per the GitHub…
pkg.go.dev API
The Go Team launched a public GET-only v1beta pkg.go.dev API for structured search and package ecosystem metadata, including packages, modules, symbols, versions, known vulnerabilities and imported-by relationships, with…
The pkg.go.dev API replaces brittle HTML scraping with an official structured interface and gives AI coding tools a cleaner evidence source. It remains v1beta,…
Evaluation Cards
The EvalEval Coalition beta-launched Evaluation Cards, an open-source reader over a normalized evaluation warehouse with card-level context and four interpretive signals: reproducibility, completeness, provenance and comparability.
Evaluation Cards is a useful antidote to treating benchmark scores as self-explanatory because it foregrounds reporting gaps and provenance. Its signals still depend on…
GitHub Agentic Workflows
GitHub released Agentic Workflows in public preview, compiling natural-language Markdown into standard GitHub Actions workflows that run coding agents for issue triage, CI analysis, documentation and other repository…
- Kingy
- 7.9 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
GitHub Agentic Workflows offers credible defense in depth through read-only defaults, sandboxing, a network firewall, safe outputs, compile-time validation and threat scanning. Those controls…
DiffusionGemma
Google DeepMind released DiffusionGemma, an experimental Apache-2.0 open-weights 25.2B mixture-of-experts model that generates text by iteratively denoising 256-token canvases in parallel.
- Kingy
- 8.3 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
DiffusionGemma is a credible architecture experiment for small-batch GPU generation, but Google’s own benchmark table shows substantial quality trade-offs against Gemma 4 on many…
Descope MCP Server
Descope launched a hosted remote MCP server for documentation search and identity-project administration, with read-only sessions by default and explicit out-of-band approval for writes.
- Kingy
- 7.7 / 10
- Demo
- Not scored yet
- YouTube
- Not scored yet
Descope MCP Server stands out for a documented read-only default, out-of-band one-time-passcode approval and a 15-minute write window. Those controls reduce silent-write risk but…
The Kingy AI Referral Index
AI assistants sent 1.3% of kingy.ai's own traffic. Published studies put the web-wide figure anywhere from under 1% to 25%. The index reconciles them, catalogued by method.
Methodology
How Kingy AI Scores Launches
Kingy AI scoring is designed for launch discovery, editorial review, creator planning, and buyer trust checks.
Kingy score
A directional editorial score for launch clarity, source quality, audience fit, pricing visibility, demo usefulness, and practical buyer or creator value.
Demo score
A signal for whether the product can be shown, tested, explained, or compared without relying on vague announcement copy.
YouTube potential
A creator-fit signal for demos, tutorials, before-and-after workflows, explainers, reviews, and audience-specific product education.
Source verification
Records distinguish verified, needs verification, and founder submitted status using official URLs, source links, last-verified dates, and correction paths.
Pricing clarity
Pricing signals favor launches with visible pricing, free-plan status, pricing pages, API access notes, or clear uncertainty when pricing is not public.
Use-case clarity
Launches are easier to evaluate when the audience, workflow, category, demo, and alternative comparison are obvious from public sources.
Founder and sponsor fit
Founder submissions and sponsor interest are routing signals only; editorial review still checks source quality, public claims, demo clarity, and usefulness.
Verification labels
Verified means enough public source evidence is present, needs verification means the record needs more checking, and founder submitted means the entry came through the submission path.
Scores are editorial signals, not scientific benchmarks, paid placements, or guarantees.
Founder submission
Submit Your AI Launch
Founders can send a launch for editorial review when the public source trail is ready to verify.
What to include
- Product name
- Website
- Launch date
- What launched
- Official source links
- Pricing page
- Demo/video
- Company info
- Category
- Public funding info
- Screenshots/media
Privacy note
Do not submit secrets, unreleased financials, private customer data, or regulated personal data.
Submit only public product, company, source, pricing, demo, and media details that Kingy AI can review without logging into private systems.
Sponsor path
Sponsor Kingy AI launch coverage
Launching an AI product that needs clear demos, creator education, and buyer trust? Sponsor a Kingy AI video or launch feature.