# AI use cases for Conch Agent (georgevibing/conch-agent) Repo: [georgevibing/conch-agent](https://github.com/georgevibing/conch-agent) | Use cases: 5 | Page: https://dedupedai.com/use-cases/georgevibing-conch-agent/ Conch is a self-hosted AI assistant app that puts Claude Code, Codex, Copilot, Gemini, Ollama and other models into one chat interface on your own computer, reachable from the web, desktop, phone and chat apps. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Chat with every model you already pay for: https://dedupedai.com/use-cases/georgevibing-conch-agent/chat-with-every-model-you-pay-for/ - Set a weekday routine that briefs you each morning: https://dedupedai.com/use-cases/georgevibing-conch-agent/weekday-morning-briefing-routine/ - Get alerted only when a watched thing changes: https://dedupedai.com/use-cases/georgevibing-conch-agent/alert-me-when-something-changes/ - Message your assistant from Telegram or your phone: https://dedupedai.com/use-cases/georgevibing-conch-agent/control-conch-from-telegram/ - Run Conch from source to develop or test changes: https://dedupedai.com/use-cases/georgevibing-conch-agent/run-conch-from-a-source-checkout/ ## How to chat with every model you already pay for with Conch Agent Repo: [georgevibing/conch-agent](https://github.com/georgevibing/conch-agent) | Difficulty: Beginner | Time: 15 minutes | Basis: From the README One local app puts your Claude, Codex, Copilot and Gemini plans in one model picker. **Who it is for:** A person who pays for several AI plans and is tired of switching between separate apps and sign-ins. **The problem:** Each AI tool has its own app, its own sign-in and its own chat history. **The outcome:** One chat app on your computer, with every connected model available from one picker. ### Steps 1. Download the app for macOS, Windows or Linux from the newest release and open it. 2. Sign in with a plan you already pay for, or paste a provider key. 3. Open the model picker and choose a provider for a new chat. 4. Ask a question, then switch models mid-chat by asking, for example "something cheaper". ### Prompt to paste into your agent ```text Using the georgevibing/conch-agent repo, connect the in the Conch app and send one test prompt. Expected output: a reply in the chat with the model name shown. Verify that the model picker lists that provider. ``` ### Caveat The models cost what your plan or key costs, and local models through Ollama or LM Studio are slower and less capable than cloud models. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/georgevibing-conch-agent/chat-with-every-model-you-pay-for/ ## How to set a weekday routine that briefs you each morning with Conch Agent Repo: [georgevibing/conch-agent](https://github.com/georgevibing/conch-agent) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Schedule a Conch routine for 7:30 each weekday and read the result in chat. **Who it is for:** A busy person who wants the day's information ready before they open a laptop. **The problem:** Checking calendars, email and tasks by hand every morning takes time before the day starts. **The outcome:** A routine that runs every weekday at 7:30 and reports back in your Conch chat. ### Steps 1. On macOS or Linux, run `curl -fsSL https://conchagent.com/install.sh | sh`. On Windows, run `irm https://conchagent.com/install.ps1 | iex` in PowerShell. 2. Open Conch and connect a provider you already pay for. 3. In chat, ask for a routine that runs every weekday at 7:30 and describes the briefing you want. 4. Open Routines in the app, confirm the schedule, and check the first result the next morning. ### Prompt to paste into your agent ```text Using the georgevibing/conch-agent repo, create a weekday routine at 7:30 in the Conch app that summarizes . Expected output: a routine listed under Routines with the 7:30 weekday schedule. Verify that the first run posts its result to the chat. ``` ### Caveat The README lists Routines and the 7:30 weekday example, but it does not describe the exact briefing format, so confirm the output on your first run. Risk, credentials: Connecting Gmail or Google Calendar gives the routine access to those accounts, so choose Read only unless the routine needs to write. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/georgevibing-conch-agent/weekday-morning-briefing-routine/ ## How to get alerted only when a watched thing changes with Conch Agent Repo: [georgevibing/conch-agent](https://github.com/georgevibing/conch-agent) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README A standing order checks now and then and tells you only when something new happens. **Who it is for:** A traveler or an operator who wants to know about a change without checking every few minutes. **The problem:** Watching a flight, a web page or a build by hand wastes time, and frequent checks become noise. **The outcome:** A standing order in your chat that checks on a schedule and reports only new changes. ### Steps 1. Install Conch with the one-line command for your system from the install section. 2. In chat, say once what to watch, for example "always tell me if a flight changes". 3. Open the check-in in the app and confirm it is listed. 4. Wait for a notice, which appears only when something new is found. ### Prompt to paste into your agent ```text Using the georgevibing/conch-agent repo, set a standing order in the Conch chat to watch . Expected output: a check-in listed in the app that reports only new changes. Verify that it stays silent when nothing has changed. ``` ### Caveat The README gives flight and page changes as examples but does not list which sources each check can read, so test your specific source first. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/georgevibing-conch-agent/alert-me-when-something-changes/ ## How to message your assistant from Telegram or your phone with Conch Agent Repo: [georgevibing/conch-agent](https://github.com/georgevibing/conch-agent) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Reach the same Conch assistant from a chat app you already use, with the same commands. **Who it is for:** A person away from their desk who wants to ask their assistant a quick question from a phone. **The problem:** The assistant lives on one computer, so you cannot reach it when you are away. **The outcome:** Your Conch assistant answers from Telegram or from the installable phone app. ### Steps 1. Install Conch on your computer with the one-line command for your system. 2. Connect Telegram from the chat apps guide in the Conch documentation. 3. Send a message from Telegram and confirm the reply comes from your Conch assistant. 4. Install the app on your phone over the private address described in the phone guide. ### Prompt to paste into your agent ```text Using the georgevibing/conch-agent repo, connect Telegram to the Conch assistant on . Expected output: a reply in Telegram from the same assistant that appears in the desktop chat. Verify that a message sent from the phone shows up in the desktop chat history. ``` ### Caveat The README says that out of the box only this computer can open Conch, so remote phone access depends on the private address setup described in the phone guide. Risk, credentials: Connected chat apps reach an assistant that can read your files and run commands, so connect only accounts you control. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/georgevibing-conch-agent/control-conch-from-telegram/ ## How to run Conch from source to develop or test changes with Conch Agent Repo: [georgevibing/conch-agent](https://github.com/georgevibing/conch-agent) | Difficulty: Advanced | Time: 45 minutes | Basis: From the README Clone the repo, install with pnpm and start the app on localhost:4317. **Who it is for:** A developer who wants to read, change or test the Conch code on their own machine. **The problem:** Installing a release does not let you run unreleased changes or edit the app. **The outcome:** A local Conch running from your checkout at http://localhost:4317. ### Steps 1. Install Node 24 or newer, then clone the repository with `git clone https://github.com/georgevibing/conch-agent`. (inferred, check the README) 2. Change into the checkout folder and run `corepack enable`. 3. Run `pnpm install`. 4. Run `pnpm start` and open http://localhost:4317 in your browser. ### Prompt to paste into your agent ```text Using the georgevibing/conch-agent repo, install dependencies and start the app from a checkout on this machine. Expected output: the app served at http://localhost:4317. Verify that the page loads and that you can send one chat message. ``` ### Caveat The README says only this computer can open Conch out of the box, and docs/SECURITY.md should be read before putting it on a network. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/georgevibing-conch-agent/run-conch-from-a-source-checkout/