We automated each tool only after a manual task kept repeating, and every archive is a sanitized copy of the real source with credentials stripped so you can read how our operation works. The full set lives in one public Drive folder.
We wrote Model Context Protocol servers so AI agents can operate our stack directly, because an agent that can read reports and file changes beats ten dashboards.
This multi-seat MCP server runs over the Amazon Ads API and covers profiles, reporting, campaign and bulk operations, and AMC queries. It is the backbone that lets our bid engines and AI workflows touch every account we manage from one interface.
Download the source archive →This slim MCP server reads and writes Sheets, Docs, and Drive through a service account, because every forecast, tracker, and client report we run lands in a spreadsheet and the automation needs the same access We have.
Download the source archive →An MCP server for OneNote runs over the Microsoft Graph API with device code auth, paired with an extract and ingest pipeline that mirrors 667 notebook pages into a full text searchable SQLite database.
Download the source archive →An MCP bridge into WhatsApp pulls chats and media so client conversations can be searched, summarized, and folded into the same operational memory as Slack and email.
Download the source archive →A working MCP bridge into the Helium 10 suite runs through a remote proxy with a scope override we set up when the native route refused to authenticate, so keyword research, listing analysis, and profit data answer inside the same AI workflows as everything else.
See the tool portfolio →Scheduled runs turn raw activity into decisions, because the daily work of 15 to 20 accounts only stays consistent when the plumbing is code.
A task engine watches Slack mentions across three workspaces, uses AI to decide accountability and extract the task, files it to Google Tasks, notifies us on Discord, and maintains a repository database of over 42,000 cross-channel work log rows that gives every new task its full account history.
Download the source archive →This is the sync layer behind our 90 day per-account forecasts, keeping projections, actuals, and pacing in one sheet per account so every weekly report is measured against a written plan rather than a feeling.
Download the source archive →A pipeline ingests meeting transcripts, extracts the action items with AI, and turns them into tracked tasks, because decisions made on calls used to evaporate and now they land in a queue with an owner.
Download the source archive →A Discord voice channel recorder captures per-user audio to separate WAV tracks for clean transcription, built for the team calls that no meeting bot could join.
Download the source archive →A client for the DataDive API drives Rank Radar keyword tracking, pulls niche and ranking juice data, and writes the tracking sheets our ranking strategy runs on, because rank movements only matter when they land next to the spend that caused them.
Download the pipeline source →A reporting client for the MerchantSpring API runs the full create, poll, and download flow, which feeds our ads team tracking sheets and the daily sales pulls behind several of the case studies on this site.
Download the pipeline source →Smaller builds that remove daily friction sit here, and the open source engines that do the daily bid work live on Tool Archive where the commit history is the audit trail.
A tab management extension suspends idle tabs while keeping chosen service tabs alive, built because a working browser with two hundred tabs is a real operating condition and not a joke.
Download the source archive →A speaking practice app records, transcribes, and scores delivery, which we built to sharpen the client facing half of the job because clear speech closes accounts as surely as clear reporting.
Download the source archive →Outside the public archives we run client specific systems that stay private by design, including hourly Prime event budget pipelines, DSP report writers, an inventory aware supply chain sync, and alerting bots, and each one is described in the case studies where its results show up.
See the results →Book a free audit call and we will walk through your campaigns live, point out where the spend is leaking, and name what we would change first.
Book a free audit call