Building the Launch Machine 🔴 Building to $100K
Want to back the stream? The «داعم» membership on YouTube is EGP 60 a month: badges, emoji, and I answer your questions first on stream. Become a supporter
The gist
The stream centers on building TalkStore, envisioned as an AI agent that takes practical actions in Shopify stores, while preparing for launch early. Luqman discusses using a landing page, waitlist, distribution campaign, and organic marketing to test merchant interest before the product is finished, while noting that sign-ups do not guarantee conversion. He iterates extensively on the landing page: documenting its visual direction, handing specifications to agents section by section, inspecting reference animations and assets, and diagnosing gaps between agent-reported changes and browser results. Some motion eventually appeared, but the design, distribution work, and teaser video still required follow-up. He also develops a multi-agent workflow for research, building, review, security, and compliance, choosing task-scoped Shopify documentation retrieval through Context7 over full ingestion without independently validating its accuracy. Other useful discussions cover automating Shopify submission audits, designing a multi-intent message classifier, resuming work after session limits, and audience questions about frontend, backend, and hosting tools. Open questions include the reliability of automatic agent-chain dispatch, completion of the reviews, the landing page’s final direction, and the distribution-automation plan.
66 moments from the stream
- 0:01:29
Why Marketing Should Start Before the Build Is Finished
He argues that developers and entrepreneurs can become absorbed in building the app or website while overlooking marketing, so he plans to prepare a pre-launch funnel early, drawing on his digital marketing and advertising experience.
- 0:06:21
Launching Before Starting: Landing Page, Waitlist, and Teaser
He outlines a plan for a landing page, waitlist, distribution funnel, teaser video, and organic promotion while the application is still at an early stage. Early attention and sign-ups could support momentum, but he acknowledges the difficulty of making a convincing teaser before the product’s appearance is settled and mentions FFmpeg as one possible technical option requiring further research.
- 0:12:17
Keeping Project Agents Current with Shopify Documentation
He considers three ways to handle Shopify’s large and changing documentation: importing all or selected parts into the project, having the research agent consult it for every task, or running targeted research before each task to stay current with Shopify requirements. The choice remains under consideration.
- 0:13:31
Diagnosing Dispatch and Reporting Gaps in an Agent Workflow
He examines a workflow where code review, security, validation, and compliance agents are supposed to follow the builder. The investigation indicates that some agents do not surface activity, while the post-build chain is not dispatched at all. The proposed remedy is to add the missing agent definitions and wire automation to trigger the post-build chain, but the excerpt does not establish that implementation was completed.
- 0:20:19
Choosing a Research Tool by Task Complexity
The speaker suggests that simple documentation research may not require a specialized research tool, while Perplexity may be more suitable when research is complex or has missing components. This is the speaker’s assessment, not a demonstrated benchmark.
- 0:22:08
Prepare-and-Dispatch Agent Lifecycle
Instructions are drafted to organize multiple agents into prepare and start phases, with a task-preparation command intended to activate them and expose their activity in the MCP feed. The speaker immediately expresses doubt about this behavior, so it remains to be verified.
- 0:23:06
Automating a Shopify Submission Compliance Audit
After a milestone is completed, the speaker notes that the compliance agent did not review the submission checklist and asks it to inspect both the checklist and codebase. The agent later appears active in the agent hub and logs its work; the audit identifies missing internal documentation justifying requested scopes for reviewer notes. The intended workflow is to rerun tracking after each build and update checklist items incrementally.
- 0:25:44
Pre-Launch Marketing Alongside Product Development
The speaker argues that product development should not happen without early marketing preparation and plans to focus on a waitlist and distribution campaign rather than paid advertising. The aim is to prepare merchant outreach alongside rapid development.
- 0:30:06
Ingest Shopify Docs or Research Them on Demand?
The speaker considers two ways to provide Shopify documentation to the project: ingest the necessary material so a research agent can review it before tasks, or keep it external and research it on demand during task preparation. Ingestion offers immediate access but can become stale as the docs change; on-demand research may stay fresher but raises questions about retrieval speed and research-tool cost. The discussion ends before a recommendation is completed, with a suggestion to narrow the scope to critical areas such as the Admin API rather than ingesting the entire documentation set.
- 0:37:17
Progressively Tuning an AI Assistant’s Behavior
The speaker argues that tuning extends beyond technical adjustment to shaping an assistant’s personality and contextual awareness around user preferences. Based on his experience, he distinguishes OpenClaw’s gradual adaptation from tuning a wrapper application, while deferring the detailed discussion to a later stage.
- 0:38:59
Task-Scoped Shopify Documentation Retrieval Instead of Full Ingestion
Because Shopify’s documentation is extensive, the chosen approach is for a research agent to retrieve task-relevant sections on demand from Context7 and pass its findings to the builder, reducing guesswork. The design moves from broad domain routing to a more granular mandatory table based on capabilities and trigger keywords, with a fallback that researches any Shopify API usage found during codebase exploration even when it is absent from the table—the table is a floor, not a ceiling. The required output also calls for precise references to relevant API operations or endpoints. The captions describe configuring this behavior, not independently validating its accuracy.
- 0:44:13
Elements of a Pre-Launch Marketing Campaign
The discussion recommends starting with organic marketing and postponing paid ads, while preparing a pre-launch distribution campaign centered on a landing page, waitlist, and teaser video. These are presented as planned areas of work, not verified outcomes from an executed campaign.
- 0:49:34
Diagnosing a React Island Integration on a Landing Page
The speaker proposes using the presence of an interactive button at the top of the page as a quick diagnostic: if it appears, the basic React island works and the likely issue is the motion library; if it does not, the React integration itself may be broken. This is presented as conditional diagnostic logic, and the captions do not record the test result.
- 0:51:40
Using OpenClaw as a Marketing Agent
A brief discussion of using OpenClaw to manage product marketing campaigns, with a positive opinion about its potential usefulness but no demonstrated results.
- 0:53:05
Impressions of Whisperflow’s Website Design
Whisperflow’s website is praised as attractive and high-end.
- 0:53:45
Reviewing the Talk Store Landing Page
A reported fix is expected to add the hero headline, eyebrow, tags, calls to action, mockup, and animation. After starting the dev server and inspecting the page, the intended animation appears absent. The discussion also stresses finishing efficiently, avoiding a generic or overly aggressive design, refining Talk Store’s style, and presenting its positioning for Shopify merchants.
- 0:55:47
Initial waitlist review and polish-task workflow
The speaker judged the design too generic and noted that expected motion, such as sequential headline-word reveals, was absent. He kept the task in review and treated it as a special case for iterative work outside the usual task-board workflow.
- 1:00:36
Cursor’s benefits and limitations for UI work
The speaker found Cursor useful for design and UI work because the browser and inspectable elements sit close to the agent, making it easier to select elements, copy snippets, and draw from other web references. He also reported terminal-sandboxing friction, voice input stopping after several exchanges, unusual paste behavior, and his opinion that its agent was capable but expensive.
- 1:04:00
Restarting a standalone Astro landing-page server
The speaker asked how to stop and restart the waitlist development server independently when the default development command launched the Shopify app instead. The displayed response identified the page as an Astro site in the landing directory and suggested managing the process on its port before restarting the dev server.
- 1:07:42
Iterative agent collaboration and visual review with Playwright MCP
The speaker chose to collaborate directly with the agent on iterative landing-page analysis and polish while leaving the task in review. He attempted to add Playwright MCP so the agent could access the page and take screenshots for visual inspection, but the captions record connection failures and do not establish that setup or analysis succeeded.
Got a question about something in this stream?
Ask in the group