Talkstore Begins Now! 🟢 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 documents the public start of building Talkstore: a proposed Shopify-specific AI assistant through which merchants could manage products, storefront work, and commerce intelligence conversationally. Its capabilities and possible expansion beyond Shopify are presented as conditional directions, not validated outcomes. Practical work covered Shopify app and development-store setup, Supabase integration and Prisma migrations, secret handling, and repeated diagnosis of session, connection, preview, and Railway deployment problems. The preview and some setup steps worked, and the service later appeared active, but several database, network, build, and deployment issues remained intermittent or unresolved in the covered segments. This led to useful lessons about starting with logs, documenting a stubborn problem, seeking a fresh diagnostic perspective, and reviewing coding-agent compromises rather than accepting a result merely because it works temporarily. The stream also presents a multi-agent command center that separates research and task preparation from implementation, code, security, and compliance review, using dependencies, an adaptive timeline, and human approval gates. Luqman felt this structure accelerated and preserved the context of the work, while still questioning whether every agent was running and visibly logged. He emphasized that authentication, secrets, and nuanced decisions continue to require human intervention, and that task count is a poor measure of difficulty. Work later shifted toward pre-launch distribution: a simple but polished Astro waitlist site kept separate from the Shopify app, with email capture, analytics, social metadata, a teaser, and a distribution campaign. The first landing-page draft had missing hero content and styling or rendering failures, so it was treated as an unfinished iteration. Other useful discussions included audience questions about using AI to explore income paths from existing expertise, SaaS unit economics, choosing between ads and SEO/GEO, Arabic-market WhatsApp automation, and targeting external markets, plus local voice dictation, content automation, and Telegram community options. Stream lag and interruptions were also investigated without a conclusive cause, while music licensing and gym culture appeared as brief secondary discussions.
121 moments from the stream
- 4:43:38
Dividing Work Between a Human and an Agent
A brief reflection on completing the task through collaboration between a human operator and an agent, acknowledging that tasks do not always work on the first attempt and pointing to the value of an agent hub with shared memory. The shared-memory explanation is incomplete on this page.
- 4:44:13
A Multi-Agent Workflow for Software Tasks
He described a task board that retains prior work so an old task can be revisited for missing or incomplete pieces. A research agent examines the codebase, tools, skills, and project status and prepares an instruction file for a builder agent, with other agents intended to review the result. He remained uncertain about the deployment behavior and wanted visibility into every agent’s activity rather than only the builder.
- 4:47:43
Uncertainty Around YouTube’s Long-Stream Duration Display
He monitored YouTube’s stream status to see whether the displayed duration would remain accurate during a long broadcast. In an earlier stream, the shown duration appeared to have been automatically shortened or reset after several hours, but he was unsure whether this related to stream health or simply the platform’s display behavior.
- 4:51:29
Diagnosing a Laggy Stream Despite Normal Speed Tests
He described a laggy stream with poor status even though speed tests appeared normal on both devices. He stopped the broadcast, changed one network-cable connection, and started another stream; this was a troubleshooting observation, not conclusive proof of the cause.
- 4:53:38
Starting the Talkstore Build and Deferring UI Polish
He moved into the practical start of building the Talkstore application. He considered making the interface visually polished immediately, then chose to revisit that later and begin with the current task first.
- 4:56:47
Prompting an agent to ask rather than assume
Luqman explains that he added a rule to an explorer agent’s system prompt: whenever human input is needed, the agent should return with questions instead of making assumptions.
- 4:58:36
Native Shopify app navigation and review risk
The discussion compares Shopify’s native navigation with a branded sidebar. Luqman accepts the native option after it is recommended as the better fit, aiming to reduce the chance of discovering a problem after the review wait, then checks the agent’s progress building the chat UI and runs the app-development workflow. He says the review may take 15–30 days, according to his estimate.
- 5:05:06
Shopify App Session Checklist
The speaker reviews recurring session setup work and connects access-token encryption and session requirements to Shopify's checklist. In his view, these requirements form a review checkpoint the project must eventually pass.
- 5:10:26
Diagnosing a Missing Prisma Session Table
After a missing Prisma session-table error appears, the speaker suspects the Shopify integration's backend rather than a paused Supabase database. He says the project remains active and the environment variables appear sound, but the excerpt does not establish a final resolution.
- 5:11:53
Agent and Human Task Categories
The speaker explains a color-coded command-center workflow: agent-executed tasks, tasks requiring back-and-forth handoffs with a human, and fully human tasks. Based on his brief experience, he found the handoff-heavy category the slowest.
- 5:13:27
Using a Research Command During Database Debugging
The speaker demonstrates a recently added command that launches a research agent while another agent handles debugging. During the exchange, he identifies a possibly missing connection parameter in the database URL and plans to add it, without confirming that the change resolves the issue.
- 5:15:58
Diagnosing a Database Connection Result
After appending a required value to the database connection URL, access became possible, but the display still showed only one sentence rather than the expected change. The discussion questioned what should appear and whether this was the chat interface, without documenting a final resolution in this segment.
- 5:18:50
Preparing and Executing Tasks with Specialized Agents
Luqman explains a workflow that separates task preparation from execution: agents for research, codebase exploration, and project management collaborate on an instruction file, then an implementation agent builds the task and code, security, and Shopify submission checks review the result. He says this structure keeps agents aware of the task and milestone and reduces his workload, while some role details remain unclear in the captions.
- 5:22:36
Verifying Agent Activity and Observability
Luqman says delegating large tasks to agents saves time, but he wants to verify that every agent is actually working because the interface or log appeared to expose activity for only some roles. He plans to inspect the agent system and its records to distinguish what is executed from what is visibly recorded.
- 5:29:43
Could Talkstore Start with Shopify and Expand Later?
Luqman thinks aloud about launching on Shopify first, explaining that expansion to other commerce platforms could become possible if the product proves itself and additional development resources become available. He presents this as a conditional future direction, not a confirmed plan. He argues that Shopify alone could be a sufficient market for a strong app with a strong marketing strategy to generate meaningful monthly recurring revenue, with later success potentially reinvested in expansion.
- 5:31:35
Talkstore’s Conversational Store-Editing Concept
Luqman describes Talkstore as an assistant that a Shopify merchant could instruct conversationally to change the storefront, such as editing a landing page after automatically duplicating the theme. He compares the idea with Shopify Sidekick, arguing that Sidekick has broad knowledge but—by his assessment—less depth in direct execution, while Talkstore targets areas including product operations, storefront work, and commerce intelligence. These are intended capabilities described during development, not final test results.
- 5:35:19
Checking App-Chat Streaming and Terminal Errors
Luqman reviews guidance for checking the streaming response, integration setup, and terminal errors. He confirms that errors are present, then says something appeared before deciding to investigate the error message further; the page shows no final diagnosis or resolution.
- 5:37:20
A Shopify-Specific AI Assistant
Luqman positions the app as a focused tool inside Shopify rather than a generic ChatGPT or coding wrapper, arguing that its specialization addresses demanding work with a potential market. After an interaction appears to run smoothly, he expresses excitement about bringing Claude into Shopify while retaining an unresolved concern about Shopify's response to the app.
- 5:41:25
Reviewing Week One and an Evolving Product Roadmap
Luqman reviews completion of the initial phase, including technical backend and settings work, then turns to a distribution track involving a landing page and waitlist. He stresses that the current board is not a final roadmap: entire milestones and features such as voice may be added later because they have not yet been discussed or recorded.
- 5:44:00
Planning a Simple, Polished Landing Page and Teaser
Luqman wants Talkstore's landing page to be extremely simple yet highly polished. He is provisionally accepting Shopify's native interface while considering further refinement or replacement, and he may separate some marketing work from the product board. He also plans to create a teaser early while acknowledging that it will require substantial time and effort.
Got a question about something in this stream?
Ask in the group