Share your best AI workflow. We could show it to 2M+ people.

“A good engineer knows when not to use AI.” I built a semi-automated workflow to summarize my live streams. With it, I can record and edit a stream in just one day of work. Not everyone has time to watch an entire stream, so creating a summary of the main moments is important for people who want to watch it later. Many tools create Shorts from videos by analyzing transcripts, but they are expensive and do not work well for visually heavy content such as gameplay. Gemini can analyze both video and audio and has a long context window, so I decided to use Google AI Studio to identify important moments. I also created scripts to automate parts of the video-editing process. Some of the scripts can be executed by an agent, and the prompts can be turned into Skills when using the API. Step-by-step: 1. I record the video with separate tracks for the microphone and background audio. I keep the microphone on the first track so the LLM does not identify only the background audio. 2. I compress and cut the video with `ffmpeg`, then extract the audio tracks. Google AI Studio has implicit video requirements: files must be under 400MB and under one hour long. The AI analyzes only one frame per second, so I can also lower the FPS to save space. I use the audio tracks later in the workflow. 3. I upload the processed videos to Google Drive, which makes them easier to use in Google AI Studio. 4. I use Gemini with temperature 1 and a high thinking level to select the important moments. I add the system prompt and specify which part of the video I am uploading: "The video is part X of the stream. Please make a structured script according to the system instructions." This helps Gemini understand what types of moments may appear in the video. The system prompt was created for gameplay presentations but can be adapted for other content. 5. I use Gemini Pro with temperature 1 and a high thinking level to find the timestamps for the selected moments. I keep timestamp generation separate from moment selection so Gemini has more thinking time for each task. Gemini Pro works better than Flash when handling time. I add the system prompt along with a copy of the response from the previous step. - As an extra check, after execution I continue the conversation with: "Check if the analysis was cut off too early (context truncation), ignoring that the video continued and generating false positives for timestamps. Check the last events especially." 6. I send the data to Blender. I create a JSON file containing the AI’s response and use a script to add the original video, the microphone track, and the background audio track. The script cuts and marks the important moments based on the JSON. Because it is not possible to send multiple audio tracks from a single video, I send the tracks separately. I also make sure to use the correct FPS, either 30 or 60. 7. I edit the video manually. The AI’s timestamps are not perfect, so I may add or remove sections, or correct a position that the AI identified incorrectly. I then render the video. 8. As an extra, I can automate standardized edits. For example, I use three different camera positions, so I add clips to three different tracks depending on the position I want. I use a script to change the position and scale of the clips on those tracks. 9. As another extra, I transcribe the audio. I mute the background audio and save only the microphone audio as an MP3, then use Whisper to transcribe it. I can use the transcript as YouTube subtitles or embed it directly into the video. I use Whisper-WebUI with Log probability Threshold -0.5, No Speech Threshold 0.5, Patience 2, and Hotwords `\u003cmy name and terms in other languages I usually use\u003e`. I use an LLM to translate the transcript into other languages. 10. Finally, I create tags and titles. I upload the transcript to Google Drive and use it in Google AI Studio with a system prompt to generate tags and titles for the video. I also use the assistant in Google AI Studio to analyze and summarize my channel for use in LLMs. This helps me choose better titles and tags.

I created a data extraction and enrichment tool using a free Google Gemini account to write the necessary Python scripts. The tool extracts contact lists based on specific search criteria and enriches the contact details through API calls to Lusha. Step-by-step: 1. I used a free Google Gemini account to write the necessary Python scripts. 2. I applied specific search criteria to extract a list of contacts. 3. I enriched the contact details using API calls through Lusha.

After about 18 months of building software with AI, I realized that reliability wasn't primarily a model problem. Bigger context windows and more clever prompts didn't fix it. What did help was treating the AI like a developer joining an existing team instead of like a chatbot. Real developers don't work from memory. They inspect production, read the documentation, check the tickets, and use proven tools. I built an environment that lets the AI do the same. The workflow is tool-agnostic, so it can be rebuilt with whatever AI client and stack you already use. Step-by-step: 1. I gave the AI a persistent task and history store that it can read from and write to. This is the core of the workflow. Mine lives behind an MCP tool, but any queryable store can work. Every architectural decision, blocker, and progress note gets written there instead of being left in the chat. 2. I open every session with a stand-up. Before writing a single line of code, the AI pulls what was in progress, what's blocked, what changed since the last session, and which architectural decisions still hold. About 30 seconds later, we're both looking at the same project. Then we build. 3. I exposed real operations as MCP tools instead of relying on "write code" prompts. I wrapped specific, tested actions—such as creating a page, defining a data model, wiring an integration, and running a migration—as tools. The AI composes these known-good building blocks into larger solutions instead of regenerating infrastructure every session. I call this wave coding, and it's the biggest reason the output stays consistent. 4. I made verification a rule: before touching anything, the AI reads the live database, API state, logs, and files. It checks ground truth first instead of making assumptions. 5. I made the chat disposable and the log canonical. If it isn't logged, it didn't happen. The task store is the single source of truth, not the conversation. The payoff is that I can stop halfway through a feature, close my laptop, and come back days later. The AI reconstructs the project from its own history, so I don't spend 20 minutes re-explaining it. Full disclosure: I built this into my own platform, WebsitePublisher.ai, which currently has 43 MCP tools and 105 integration building blocks. It's delivered as an add-on that plugs into the AI client I already use over MCP, so there's no new app to learn. Nothing here is locked to that platform, though: the workflow itself works with any MCP client and any store the AI can query. I'm curious whether anyone else is running their AI this way or solving the amnesia problem from a different angle.

My AI digest looked perfect and was quietly wrong. What actually fixed it. Every run succeeded. Every dashboard was green. And the content was still wrong. My digest invented "AI Moat Brief", a newsletter that does not exist. It reported scan counts nobody measured. It resurfaced week-old stories as fresh headlines. Here is what broke, and what fixed it. The sorting used to happen in my head: skimmed subject lines, unopened tabs, quiet guilt. The Signal is one email at 08:00: a single Make scenario calling Claude Sonnet through OpenRouter. It reads the last 24 hours of my RSS feeds and newsletters, keeps what touches what I am actively building plus the domains I need to stay current in, and arrives in the language I actually think in. Core items end with what it means for my work. Five to ten minutes, and I know where to go deep today. Structurally it looks like this, minus the content, rendered in English for this post (Image 1). No real edition is shown; section names and sample lines are illustrative. The dangerous failures were never pipeline failures. They were content failures, and the cause is structural: an LLM summarizing newsletters that already summarize primary sources is third hand by construction. Every hop strips attribution and adds confidence, and when data goes missing the model fills the gap the way LLMs do: fluently. Valid HTML, confident tone, green pipeline, wrong content. Image 2 is that whole failure class in one frame. Three rules closed the gaps I caught, all live in production: Step-by-step: 1. Verbatim or nothing. A source name is copied character for character, and a link exists only if that exact URL is in the input. The model copies; it never composes. 2. The model never generates metrics. Every count the report shows is injected by the pipeline after the model returns. 3. Recycled news gets demoted. A recap of recaps gets one line at most, and is dropped when the underlying story falls outside the collection window. Rules 1 and 3 lean on the prompt, and that is why the counters exist. The pipeline writes a hidden HTML comment into every email it sends: items, links, urls, cost, finish status. That line caught what I could not see. In one run, the published-links counter and the leftover-urls counter read 45 and 435: the only sign a new cleanup step was a silent no-op. Another morning the model stopped at 15,999 tokens against a 16,000 cap, one token from an email cut off mid-sentence. On the morning I wrote this they agreed, 21 links and 21 urls, and boring is the goal. Image 3 is that morning's actual comment, with the same two counters from the no-op run. The run itself has a dead man's switch on Healthchecks.io, so a missing 08:00 email reaches me before I notice. Honest limits: the $0.31 per report is a fresh measurement I am still validating, and I have not proven these rules hold as the source set scales. There is more behind every part of this; I would rather share it where it is wanted. Ask and I will put it in the comments: the three rules in full, the exact cost and what drives it, what this replaced in my day, how it compares to what is on the market, or the ugliest of the 15 documented bugs. I am sharing this because I doubt I am the only one building fragile things behind the scenes, and monitoring text is harder than monitoring uptime. What content-level checks do you run on LLM output, the kind pipeline monitoring cannot see? Real thresholds and embarrassing failures especially welcome.

AI agents are powerful, but they rarely reach the apps that run daily life. Amazon, Uber, Instacart, Walmart, and DoorDash expose little or no public API access, while simulated input through desktop automation or ADB can leave software fingerprints that anti-bot systems flag. The alternative is to give the agent an arm and an eye and let it operate a phone. The screen becomes the API: a camera watches a real phone, and a robotic stylus taps it. From the phone’s perspective, the input is indistinguishable from a human finger. Nothing needs to be installed, and there is no OAuth setup. Hardware is slower than an API call—each action takes a few seconds—but it can reach virtually any app. Step-by-step: 1. I message the agent like a friend. It has its own phone and its own chat account. 2. The screen lights up, the runtime wakes the agent, and it unlocks the phone and reads my message. 3. The agent works out what I want, opens the right app, and operates it by hand using taps, swipes, and scrolls. 4. If an action involves spending money, the agent pauses and asks for my confirmation. 5. It finishes the task, replies with the result, saves what it learned, and goes back to sleep.

I’m learning Chinese online and record all my lessons. I asked Claude Code to transcribe the recordings and walked through the process of using the Whisper API with its guidance. Then I asked Claude Code to analyze the transcripts and identify mistakes I could fix that would make a big difference. It found that my teachers had not corrected several phrases I was repeating. Claude Code taught me which phrases to practice, and I’m now making fewer mistakes. Step-by-step: 1. I recorded all of my online Chinese lessons. 2. I asked Claude Code to guide me through transcribing the recordings with the Whisper API. 3. I asked Claude Code to analyze the transcripts for mistakes that would make a significant difference if corrected. 4. I reviewed the repeated phrases my teachers had not corrected. 5. I practiced the phrases Claude Code identified, which helped me make fewer mistakes.

My husband and I run a short-term rental business managing two properties we own in Harpers Ferry, WV. A major part of the job is using PriceLabs, a dynamic pricing and revenue management tool that tracks our properties’ performance against the market, including occupancy, prices, and revenue, and recommends price changes when needed. Previously, we had to log in and review multiple dashboards to make informed decisions. Now, every two days we receive a concise morning briefing that summarizes how the month is going and what needs attention. It’s one of the first things I read in the morning, so I know what to adjust in real time. Step-by-step: 1. I connected the data sources Claude needs through MCP connectors: PriceLabs for reservations, pricing, and market data, and Gmail for drafting the output. 2. I wrote the instruction prompt Claude runs each time. This took the most time to develop. 3. I defined the format and length: a short, numbers-first brief of about 200 words, beginning with “How This Month Is Going” and “What Needs Attention,” with no filler or pep-talk tone. 4. I solved delivery by having the routine draft the briefing as an email with a recognizable subject prefix, such as “Iconic Chalet Briefing — Mon, Aug 10.” Because Claude’s Gmail connector can create drafts but not send emails, a separate time-triggered Google Apps Script watches for drafts with that subject prefix and sends them automatically. 5. I registered the workflow as a scheduled Routine. That combination makes the process feel automatic from end to end.

I regularly use ChatGPT for lesson planning and materials development with students ranging from EFL learners to UK undergraduates. I keep a separate anonymous thread for each student, so GPT can track the context and suggest the best direction for the next session. I also contribute my experience as a tutor while using AI to create fresh worksheets based on typical materials from various UK exam boards. Today, I had a successful interview for a lecturer role. Gemini helped me design and plan my demo seminar, and it looks like I’ve landed the job. Step-by-step: 1. I create a separate anonymous thread for each student. 2. I use ChatGPT to track each student’s context and suggest the best direction for the next session. 3. I combine AI-generated ideas with my experience as a tutor when planning lessons and developing materials. 4. I use typical materials from various UK exam boards as the basis for creating fresh worksheets. 5. For my lecturer interview, I used Gemini to design and plan my demo seminar. 6. I had a successful interview, and it looks like I’ve landed the job.

I often lose track of movies and TV shows I want to watch, especially when they are announced months before release. To solve this, I use a free API from watchmode.com. I asked Hermes Agent—although ChatGPT Work could probably do the same—to create a scheduled task that checks the API once a week and notifies me when a title becomes available to stream. I also gave it a simple instruction: whenever I send an IMDb link for a movie, TV show, or specific season, it should automatically add it to my watchlist. Since I prefer to binge-watch, it only notifies me when an entire season is available. The next step is to make it automatically start tracking the following season once the current one has been fully released. Step-by-step: 1. I use the free API from watchmode.com to check streaming availability. 2. I asked Hermes Agent to create a scheduled task that checks the API once a week. 3. I configured the task to notify me when a movie or TV show becomes available to stream. 4. I instructed it to add a movie, TV show, or specific season to my watchlist whenever I send an IMDb link. 5. I configured notifications for TV shows to wait until the entire season is available. 6. My next step is to have it automatically start tracking the following season after the current one has been fully released.

I asked Claude to build me an artifact that works as my own personal book tracker. It includes separate sections for books I want to read, books I’m currently reading, and books I’ve read. I also asked Claude to collect specific data points about each book and create an insights tab that can gather information about my likes and dislikes over time, then make future recommendations. I personally dislike Goodreads’ UI and don’t use its social media features, so I wanted a personalized alternative housed within Claude alongside all my other workflows—a one-stop shop. Step-by-step: 1. I asked Claude to build an artifact for tracking my books. 2. I organized the tracker into books I want to read, books I’m currently reading, and books I’ve read. 3. I specified the data points I wanted to collect for each book. 4. I asked Claude to add an insights tab to gather data about my likes and dislikes. 5. I set it up to use those insights for future recommendations and keep the entire workflow within Claude.

My pool water looks great, and I’ve taken over the weekly water-care routine myself after years of using a pool service. I first gave Claude the relevant information about my in-ground pool setup. Now, each week, I check the water’s chlorine, pH, alkalinity, and other chemical levels, then enter the results into Claude. I often take a water sample to a local pool store, where they provide a printed analysis. I snap a photo of the analysis with my phone and upload it to the Claude chat. Claude tells me which chemicals to add and when. Over time, the water has stayed balanced and looks great, while I’ve saved a lot of money on pool service and avoided wasting chemicals. This workflow has paid for itself many times over through this one use case. Step-by-step: 1. I gave Claude the relevant information about my in-ground pool setup. 2. I checked the pool water weekly, including chlorine, pH, alkalinity, and other chemical levels. 3. I either entered the results into Claude or took a water sample to a local pool store for a printed analysis. 4. I photographed the printed analysis with my phone and uploaded it to the Claude chat. 5. I followed Claude’s guidance on which chemicals to add and when. 6. I repeated the process over time to keep the water balanced and avoid wasting chemicals.

I’ve automated 90% of my payroll work at an engineering firm. I use Claude Code to write Python scripts that pull timesheets from our ERP, Deltek Ajera. I check the data for oddities and, once everything looks right, push the timesheets into our HRIS, Paycom, through its API. I also have a script that parses and compares the preliminary payroll register with the timesheet data. Once I’m satisfied with the results, I can submit payroll. After payroll, my scripts create all the benefits upload files I used to create by hand. The spreadsheets I relied on for years are no longer necessary. Do I even need a spreadsheet anymore? Step-by-step: 1. I use Claude Code to write Python scripts for the payroll process. 2. I pull timesheets from Deltek Ajera, our ERP. 3. I check the timesheets for oddities. 4. Once the data looks right, I push the timesheets into Paycom, our HRIS, through its API. 5. I parse and compare the preliminary payroll register with the timesheet data. 6. Once I’m satisfied with the comparison, I submit payroll. 7. After payroll, I run scripts that create the benefits upload files I previously created by hand.

Create an immersive 3D website for artists with an AI avatar that welcomes visitors, answers questions about the artist and artwork in real time, and presents the work throughout the virtual gallery. Step-by-step: 1. Build or download a 3D environment using Claude Code or another suitable development tool. 2. Upload the artworks into the 3D environment. 3. Create a digital avatar, optionally using Adobe Mixamo (free), and import it into the scene. 4. Connect the avatar to an AI language model so it can interact naturally with visitors and answer questions about the artist and the artwork. 5. Give the avatar the ability to point to, manipulate, or present artworks during conversations. 6. Write a detailed artist biography and a description of each artwork. Use this content as the AI avatar’s knowledge base so it can provide accurate, engaging responses. 7. Use open-source voice technologies through Pinokio to enable realistic speech synthesis and voice interaction. 8. Arrange the artworks throughout the 3D environment and optionally display videos on the gallery walls. 9. Configure the audio system so multiple videos do not play sound simultaneously. Audio should activate only when a visitor approaches or interacts with a specific artwork or video. 10. Configure collision detection so the avatar cannot walk through walls or other obstacles. For example, stop the avatar whenever it encounters an obstacle higher than approximately 60 cm (24 inches). 11. Optimize the experience for the web and deploy the website using FileZilla and OVH hosting. 12. Pay special attention to iPhone and iPad compatibility. Apple devices are more sensitive to large assets and memory usage, so it may be necessary to create separate optimized versions for iOS and Android, using compressed models, textures, and videos to ensure smooth performance.

I built an Awish workflow that handles a support ticket before anyone on the team opens it. I realized support tickets were not just taking time to answer. Someone still had to understand the problem, decide how urgent it was, search the documentation, route it to the right person, and prepare a response. I wanted that first layer of support work to happen automatically. I opened Awish and wrote: “Whenever a new support ticket comes in, understand the issue, determine its urgency, check our documentation, prepare a response, create a Jira issue if it looks like a product bug, and escalate anything important to the team before taking customer-facing action.” Awish understood the request, planned the workflow, and showed me which applications it needed. Step-by-step: 1. I described the complete support process I wanted in the Awish chat. 2. Awish planned the workflow and selected Zendesk, Notion, Jira, and Slack for the required steps. 3. I connected my accounts and approved the automation plan. 4. When a new Zendesk ticket arrives, the workflow identifies the issue type, urgency, and customer intent. 5. It checks the relevant Notion documentation and prepares a response based on the available information. 6. If the issue looks like a product bug, it creates a Jira ticket with the customer context already attached. 7. Urgent or sensitive cases are escalated to the team instead of being handled automatically. 8. I connected WhatsApp from the Awish chat. I now receive workflow updates there and can manage the automation without opening Awish. Customer-facing actions still stay under my control when approval is needed.

Post-production for each of my podcast episodes used to be a significant task, even when an episode was audio-only. I’ve now built a workflow with Claude and Descript to automate the production process. I clean up the original recording in Descript, export the transcript to Claude, and run a single command to generate the remaining assets. This includes show notes for Hello Audio, opening and closing scripts written in my voice using anti-AI files to avoid a robotic tone, audio clips and audiograms for social platforms through a custom Claude skill, teaser posts for each audiogram on LinkedIn, and launch-day posts for LinkedIn and Substack, where my audience is. I’m sure there’s a way to streamline the workflow further to include distribution of the posts, audiograms, and episodes. It’s a work in progress 😀 Step-by-step: 1. I clean up the original podcast recording in Descript. 2. I export the transcript from Descript to Claude. 3. I run a single command in Claude to produce show notes for Hello Audio. 4. I generate opening and closing scripts for each episode using my voice and anti-AI files to avoid a robotic tone. 5. I use a custom Claude skill to automate the production of audio clips and audiograms for social platforms. 6. I create teaser posts for each audiogram on LinkedIn. 7. I prepare the main LinkedIn and Substack posts for each episode’s launch day. 8. I’m continuing to look for a way to streamline distribution of the posts, audiograms, and episodes.

I use Copilot to build an agent that critically reviews my internal audit work. I only need to upload a format. 1. I upload the format to Copilot. 2. I use Copilot to build an agent that critically reviews my internal audit work.

I’m a freelancer on Fiverr and Upwork. I use ClickUp to manage enquiries, projects, and deadlines, while Slack is my team communication tool. When I receive a high volume of enquiries, I sometimes miss replies to older orders. That can also cause me to overlook important messages and hurt my responsiveness. To address this, I created a workflow in Make.com. Every four hours, I receive alerts about relevant messages. Urgent priorities are marked in Slack, and pending follow-ups that need attention are highlighted. This helps me quickly reply to all relevant messages. Step-by-step: 1. I manage all Fiverr and Upwork enquiries, projects, and deadlines in ClickUp. 2. I use Slack for team communication. 3. I connected the workflow in Make.com to send alerts every four hours. 4. I mark urgent priorities in Slack. 5. I highlight pending follow-ups that require attention. 6. I use the alerts and highlighted messages to reply to relevant enquiries and orders promptly.

AI agents can generate impressive work quickly, but the agent that created something is not necessarily the best judge of whether it is correct, complete, secure, or ready to move forward. Without an independent review step, mistakes can compound as later stages build on work that was never properly validated. I created a system of specialized AI review agents that act as quality gates between stages of work. Instead of letting the agent that performed the work decide whether it is finished, a separate reviewer evaluates the output against explicit criteria and makes a gate decision: PASS or NEEDS REVISION. Different reviewers focus on different dimensions. In my software development workflow, I use reviewers for implementation fidelity, code quality, security, performance, and specification compliance. A feature does not advance until the required reviewers have passed it. Step-by-step: 1. Define what “good” means before the work starts. Give reviewers an explicit source of truth, such as a specification, plan, acceptance criteria, coding standards, security rules, or quality rubric. 2. Separate execution from evaluation. The agent that performs the work should not be the only agent deciding whether that work is acceptable. 3. Create specialized reviewers for important quality dimensions. For software, this might include implementation, code quality, security, performance, and specification reviewers. The same pattern can be used for research, writing, factuality, compliance, financial analysis, or brand review. 4. Run the appropriate reviewers when a stage is complete. Each reviewer independently inspects the work from its assigned perspective and actively looks for reasons it should not advance. 5. Require an explicit gate decision. A reviewer must return either PASS or NEEDS REVISION, along with concrete findings and recommended fixes. In my workflow, reviewers can block progression for issues such as missing tests, even when the underlying implementation appears correct. 6. Route failed work back to the appropriate agent. The worker fixes the identified problems and submits the work for review again. 7. Advance only after the required gates pass. Later stages should not build on work that still has unresolved review findings. 8. Keep humans at consequential decision points. AI reviewers can determine whether work satisfies their assigned criteria, but important actions such as merging, deploying, publishing, or otherwise committing the result can remain human decisions. Instead of treating AI-generated work as complete simply because an agent produced it, I create a controlled loop: Build → Review → Fix → Re-review → Pass → Advance The result is a more reliable workflow where specialized agents perform the work, independent agents challenge it, and errors are caught before they propagate into later stages.

My car recently lost a small part near the bonnet. I didn’t know what it was called, and even the official workshop couldn’t help because the model was no longer in production. I gave ChatGPT my car details and a photo showing where the part was missing. It identified the part, helped me cross-check the specifications and compatible part numbers, and found sellers that had it available. I found the right part and had it delivered within a week. Step-by-step: 1. I shared my car model and a photo of where the part was missing. 2. I used ChatGPT to identify the part and its specifications. 3. I cross-checked compatible part numbers and alternatives. 4. I searched for sellers carrying the correct part. 5. I compared the options and ordered the right one.

For the start of the new school year, I wanted to use the current eastern Washington wildfires near Spokane with my two sixth-grade science classes. I wanted an AI tool to create a website I could upload to my hosting service. The site needed to provide daily updates from verified news sources, along with air quality data comparing Spokane in eastern Washington, Kirkland near Seattle, and Chimacum, where we live, on the western side of the state. I also wanted to include a simulation where students could change different variables to see how they affect the direction of the smoke and air quality. I could not decide which AI service to use, so I entered the same prompt into my personal paid Claude and Gemini Spark accounts, as well as my free teacher ChatGPT and Claude accounts, to compare which one created the best site for students. Here is the prompt I used: "Create a dynamic website with information of the Spokane, WA 2026 wildfires. Information include only verified facts from credible news sources. Information needs to be updated from when the largest fire started until it is completely put out. At that point, this website is complete and done. Include a page with air quality data. Provide air quality data on a daily basis from the onset of the fire to the end of the fire. Include Spokane, Kirkland, and Chimacum to compare the air quality from eastern WA near the fire, mid WA, and western WA. Include factors that explain the air quality such as winds, inversion, altitude of smoke and any other relevant data that students can view to answer questions. Provide a way for students to click on different variables to see how different factors affect the fire, direction of smoke, and air quality. Use the /inversa-design skill to make the website." Here is a portal page with links to all four sites: https://educatoral.com/2026SpokaneFire/ Step-by-step: 1. I identified the eastern Washington wildfires near Spokane as a topic for my two sixth-grade science classes. 2. I defined the website requirements: verified news updates, daily air quality data, comparisons among Spokane, Kirkland, and Chimacum, and a student-facing smoke and air quality simulation. 3. I used the same prompt with my paid Claude and Gemini Spark accounts and my free teacher ChatGPT and Claude accounts. 4. I compared the sites created by the four AI accounts to determine which one would be best to use with students. 5. I collected links to all four sites on this portal page: https://educatoral.com/2026SpokaneFire/