Week 27 Recap: A $50 ChatGPT Ads Test, Claude Code Inside Android Studio, and Ads in AI Answers
November 8, 2026
1,188 upvotes on the week 45 board, Tire Kicker on top with 51, SummaGraph first in the free section with 30, a $50 test plan for ChatGPT Ads before the $500 credit offer ends, Claude Code and Android Studio tested on a real app, and three AI news blocks checked against the source. Week 27 on Nick Launches.
Your weekly roundup of fresh launches shaping how we work and build.
This week is about where your next users come from. ChatGPT and Google now sell the spot under the AI answer, and OpenAI will match $500 of spend with $500 of credit until December 31, so there is a guide to testing it on $50 first. And Android Studio now runs Claude Code, Codex and 40 other agents, so the dev guide sets that up and runs it on a real app.
The week 45 board cast 1,188 upvotes, up from 1,118. Tire Kicker stayed on top with 51, and SummaGraph led the free section with 30. Every landing page in this issue got a hand check before making the list.
In this issue:
- The week 45 board: Tire Kicker on top with 51, eight featured launches, SummaGraph first in the free section with 30
- New launch guide: test ChatGPT Ads on a $50 budget before the $500 credit offer ends
- New dev guide: Claude Code inside Android Studio, and the Android CLI plugin run on a real app
- Ads arrive in AI answers: OpenAI's credit, Google's AI Mode in ten more countries, and beehiiv's crawler numbers
- The week in models and prices: ChatGPT and Codex share one allowance from December, a shorter GPT-6, Kimi's weights
- Agents move to your phone and your IDE: DigitalOcean Cuttlefish, Perplexity Teammates, GitHub Copilot connectors
Featured Launches 🚀
The Pro launches on this week's board, ordered as the board had them at the end of Sunday, with each one's real board position, and every landing page checked by hand. Eight this week: five carried over from earlier weeks and still climbing, and three new.
-
Tire Kicker: 51 upvotes, #1 on the board. Still climbing: it had 37 last issue.
Three AI shoppers try to buy from your Shopify store, the three ways real ones arrive, and you see where each one
got stuck. Free once a month per store. Featured through November 10. -
DevFeed: 23 upvotes, #2 on the board, up from 9 in week 25. It was on last week's board
too, and I missed it in the last issue. A free feed of developer news, tutorials and releases you can read without an
account. Featured through November 20. -
DisplaySwitcher: 18 upvotes, #3 on the board, up from 9. Switch your monitor's
input from a button on screen and move text and files between your computers. Featured through November 11. -
JEV Ultrafast Agent: 13 upvotes, #4 on the board, up from 9. The free,
open-source browser agent for Chrome and Edge that works with the model you pick. Featured through November 11. -
Inkbound: 11 upvotes, #5 on the board, new this week. A writing app in the browser for
novelists: chapters, characters and the world of the book in one Codex, a style aid that flags repetition and weak
verbs, a timeline you arrange by dragging, version history, and export to Word, PDF or ePub. Free for two projects;
Premium is a one-time $20 at launch, and adds AI that finds your characters in the manuscript, corrected voice
dictation and a tension map for romantic arcs. It never writes for you. Featured through November 15. -
AEO Footnotes: 7 upvotes, #6 on the board, up from 5. Sourced writing on AI
visibility, every article with numbered footnotes. Featured through November 28. Disclosure: AEO Footnotes is my own
site. -
SilverSafe AI: 6 upvotes, #7 on the board, live since Saturday. Scam protection for
seniors and the people who look out for them: call or text SAM about a suspicious call, text or email and get guidance
on the warning signs. The first two scam checks are free, and the Essential Plus monthly plan starts with a 7-day
free trial. Featured through November 20. -
Prompt Generator: 3 upvotes, #9 on the board, live since Sunday. Pick the AI tool
you use, ChatGPT, Claude, Gemini, Midjourney, Nano Banana or Veo among them, describe a rough idea, and get a prompt
in the syntax that tool reads. 10 prompts a day free without an account. Featured through December 8. Disclosure:
Prompt Generator is my own product.
Launch Weekly Winners 🏆
The most upvoted free launches of the week. No paid placement here, just upvotes. All five went live on Monday, and three of them tied on 28.
-
SummaGraph: 30 upvotes. Paste text, notes or a long article and get an infographic designed around the content rather than dropped into a template, with 22 visual styles, 20 layouts and more than 10 output languages. Free starter credits, then one-time credit packs.
-
Boothstrip: 29 upvotes. A free photo booth in the browser: four photos, a vintage,
sepia or black and white filter, grain and a date stamp, out as the tall four-frame strip of an old mall booth. No
sign-up, no watermark, and the photos never leave your device. -
TrackTimer: 28 upvotes. Time tracking with AI for freelancers, consultants and
agencies: capture the work, bill the client, and see what every project earns. Free during early access. -
ReddMatch: 28 upvotes. Finds Reddit conversations where someone needs what you
sell, drafts a reply that fits the thread, and leaves the posting to you. The first 3 leads are free. -
Notion Elevation: 28 upvotes. Notion frameworks and templates built around
systems rather than single pages, for creators and knowledge workers who want a setup that outlasts the tool.
Four more finished on 27: Tool List, OffBooks.ai, EZPZ File and Fleden. Then Sky Launch on 26, and Holosign, Math Charts, IntentLoom and WheySearch on 25.
Launch Builders to Watch 👀
Builders who made this week what it was. Worth a follow.
- @abhi16_93 (opens in a new tab): DevFeed, 23 upvotes and #2 on the board in its third week.
- @NeuriseIO (opens in a new tab): JEV Ultrafast Agent, up to 13 and still featured at #4.
- @silversafeai (opens in a new tab): SilverSafe AI, scam checks for seniors by call or text, featured at #7.
- @aiden_ok (opens in a new tab): SummaGraph, 30 upvotes and first in the free section.
- @pfista (opens in a new tab): TrackTimer, 28 upvotes for time tracking that bills the client.
- @getreddmatch (opens in a new tab): ReddMatch, 28 upvotes for Reddit leads you reply to yourself.
- @Guronnimo (opens in a new tab): Tool List on 27, Sky Launch on 26 and Fast Launch on 23, three launch boards in one week.
- @offbooksai (opens in a new tab): OffBooks.ai, 27 upvotes for books, profit and cash flow in one place.
- @xsawyerai (opens in a new tab): Holosign, 25 upvotes for e-signatures with unlimited users.
- @michaelyinplus (opens in a new tab): IntentLoom, 25 upvotes for buying signals on Reddit.
What's happening on Launch 📊
This week
- 📥 425 products submitted, up from 374
- 🚀 84 went live in their launch window, down from 93, 4 of them featured
- ⬆️ 1,188 upvotes cast, up from 1,118
- 👤 511 new builders joined, 7,405 in the directory
- 📬 483 new subscribers on this newsletter, 5,026 on the list
The bigger picture
- 🟢 2,604 live products on the directory, up from 2,418
- 🔗 Domain rating holds at 77
Test ChatGPT Ads on $50 💸
ChatGPT shows ads under its answers now, to logged-in adults on the Free and Go plans. OpenAI's ads page offers $500 of ad credit when you spend $500, and the offer ends December 31. That is a lot to put into a channel you have never tried.
The guide is the step before it: a $50 campaign in OpenAI Ads Manager that tells you whether people who ask ChatGPT about your problem click on your answer, and whether they do anything after.
What's inside:
- What the offer says, and what OpenAI's pages leave open: who counts as new, and when the credit expires
- One Clicks campaign, one ad group, two ads, with context hints written as the questions buyers ask
- OpenAI's limits: a title of 16 to 24 characters recommended, ad copy of 32 to 48
- The $50 arithmetic: the $25 daily minimum can spend $50 in one day, and $50 at a $3 to $5 bid is 10 to 16 clicks
- What to read in the report, and the stop rules to write down before you start
It is written from OpenAI's help center and developer docs. No ad account was opened and no money was spent for it, and it says so at the top.
Claude Code inside Android Studio 🤖
Android Studio Rabbit 2 went stable on October 28 with Bring Your Own Agent: more than 40 coding agents, Claude Agent, Codex, Junie and Antigravity among them, run inside the IDE through the open Agent Client Protocol, with access to builds, the emulator and Logcat. The same day, Google shipped the Android CLI as a plugin for Claude Code and Codex.
The guide sets up both on a Mac and runs them on a new app, from an empty template to a screen on the emulator that says "Hello from Claude Code", and it says plainly which parts were run and which are from Google's docs.
What's inside:
- Claude Agent added from Android Studio's agent marketplace, which runs it with npx and needed no new sign-in here
- The IDE tools the agent gets: Gradle builds, Logcat, screenshots, deploy and Compose previews, by name
- The Android CLI plugin for Claude Code, run headless: an empty template to a screen on the emulator in 7.9 minutes
- The four things that broke on a real Mac, from a failing
android buildto a missing emulator skin, and the fixes - One honest flag: Google calls the feature stable, and the IDE still labels it preview
Codex, Junie and Antigravity were not run inside the IDE, and the guide marks those parts as written from the docs.
Ads arrive in AI answers 📢
Three things happened in the same two weeks, and together they say where AI search is going for anyone with a small budget.
- OpenAI wants small advertisers before the year ends. $500 of ad credit when you spend $500, through December 31, for eligible new advertisers. The US daily minimum is $25, and OpenAI's help center suggests a starting bid of $3 to $5 per click. From November 9, new click campaigns can use Target CPC, an average you aim for rather than a ceiling. The guide above is how to test it without committing $500.
- Google is putting ads in AI Mode in ten more countries. On October 28 it named Australia, Canada, India, Kenya, Malaysia, New Zealand, Nigeria, Pakistan, the Philippines and Singapore, with Brazil, Mexico and the UK planned for the fourth quarter. In the US it is launching ads in follow-up questions. Coverage called that a test; Google's post says it is launching.
- beehiiv counted what AI crawlers give back. In a post dated October 25, beehiiv said AI crawlers made 448 million requests to its publications in the past 30 days and generated just 58,000 direct referrals, about 0.2% of the referral traffic. That is about 7,700 requests for every visit sent back.
For a builder: AI answers read your pages and send back little traffic, and now the spot under the answer is for sale. That cuts both ways. If your buyers ask ChatGPT the question your product answers, a $50 test is a cheap way to find out whether that spot sells. If they do not, the money is better spent where they already are.
The week in models and prices 🧠
- ChatGPT and Codex will share one allowance. OpenAI's help article says that starting in December, the usage allowance on Plus, Pro and Business is shared across ChatGPT and Codex, as the agent features of Work move into the main ChatGPT app. OpenAI has not published the new allowances. Today, Pro plans have no five-hour limit in Work and Codex, and Plus and Business do.
- GPT-6 in ChatGPT got shorter and less flattering. OpenAI's model release notes, dated October 29: GPT-6 in Chat now gives more concise replies with less flattering language. That applies to Free, Go and logged-out users. On Plus and Pro it only changes the fallback model ChatGPT switches to after the GPT-6 limit; the main GPT-6 there is unchanged.
- Kimi K3.1 Pro's weights are out. Last issue they were promised by November 4 and not yet on Hugging Face. They are there now, at moonshotai/Kimi-K3.1-Pro: 2.8 trillion parameters, 104 billion active, a 1 million token context. The license is permissive with two conditions: a model-as-a-service business with more than $20 million of revenue in a year needs a separate agreement with Moonshot, and a product with more than 100 million monthly users or $20 million of monthly revenue has to show "Kimi K3.1-Pro" in its interface.
Three items from this week's newsletters are not here. DeepSeek did not double V4 Pro's price: its pricing page still shows $0.66 in and $1.98 out off-peak, and peak hours have always cost double. Mistral Large 4's weights and Claude's Ultra plan without the five-hour limit were both in last issue.
For a builder: if you code with Codex on a ChatGPT plan, December is when your chats and your agent runs may start drawing on the same allowance, and OpenAI has not said how much each one counts. Check Settings, Usage once the change lands, before you plan a long agent run.
Agents move to your phone and your IDE 📱
- DigitalOcean Cuttlefish, November 5. A native iPhone app for its Managed Agents platform, on the App Store for Managed Agents customers. You follow a live session, read the full command, file path or diff before you approve or deny a tool call, and steer, pause or stop the run. The agent runs in DigitalOcean's cloud, so locking the phone does not stop it. It works with Claude Code, Codex CLI, OpenCode, OpenHands, Hermes, LangGraph and CrewAI. Managed Agents itself is in public preview, which DigitalOcean says is not supported for production work.
- Perplexity Teammates, November 4. Agents for work that takes weeks: each runs in its own environment, coordinates Computer sessions and other teammates, and can sit in a Slack channel where colleagues hand it work. It is a research preview on the web for individual Max subscribers, $200 a month or $2,000 a year, with iOS to follow. Enterprise plans are not included by default, and runs draw on Max's monthly credits.
- Android Studio opens to 40+ agents. Covered above, with the guide.
- GitHub Copilot connectors, October 28. GitHub-verified MCP servers you connect once and use across the Copilot app, Copilot CLI, VS Code, Copilot in the cloud and GitHub Mobile. 26 services at launch, including Slack, Figma, Notion, Sentry, Linear, Cloudflare and Vercel. Public preview on all Copilot plans; on Business and Enterprise an admin has to turn on the Editor preview features policy first.
For a builder: the pattern this week is approval moving to wherever you are. The agent runs somewhere else, and you answer "yes, run that" from a phone, a Slack channel or the IDE you already had open. Before you wire one up, decide which actions it may take alone and which wait for you.
Now on six platforms 📣
X stays the daily feed. The other accounts get the SEO and launch content as short video and carousels.
- X: @nicklaunches (opens in a new tab), the daily board
- TikTok: @nicklaunches (opens in a new tab), one tactic per clip
- YouTube: Nick Launches (opens in a new tab), Shorts and walkthroughs
- Instagram: @nicklaunches (opens in a new tab), the carousels
- LinkedIn: Nick Launches (opens in a new tab), the longer write-ups
- Facebook: Nick Launches (opens in a new tab), the weekly board
Pick the one you actually open.
Quick ask 🙏
1,188 upvotes, eight featured launches, and three free launches tied on 28 behind SummaGraph and Boothstrip.
If something in this issue deserves eyes, open its page and upvote it. Ten seconds, real difference. Tire Kicker is featured until November 10, DisplaySwitcher and JEV Ultrafast Agent until November 11, Inkbound until November 15, DevFeed and SilverSafe AI until November 20, AEO Footnotes until November 28, and Prompt Generator until December 8.
Thinking about ChatGPT Ads? Read the $50 test first. It costs nothing to read and it might save you $450.
Thanks for being part of Nick Launches. If you want the daily behind the scenes, including which launches climb and why, follow me on X: @nicklaunches (opens in a new tab).
Building something? Submit it. Listing is free and the backlink is permanent.
27 weeks down. See you Monday,
Nick