# AI use cases for DeepSeek Bot (rlaope/deepseek-bot) Repo: [rlaope/deepseek-bot](https://github.com/rlaope/deepseek-bot) | Use cases: 4 | Page: https://dedupedai.com/use-cases/rlaope-deepseek-bot/ A self-hosted AI teammate, built on DeepSeek Harness, that lives in Slack, Discord and Telegram and also runs as a terminal app (dsh). This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Add a remembering AI teammate to a Slack channel: https://dedupedai.com/use-cases/rlaope-deepseek-bot/team-channel-memory/ - Get docx, xlsx or pptx reports drafted in chat: https://dedupedai.com/use-cases/rlaope-deepseek-bot/draft-report-files-in-chat/ - Set a watch or reminder that the bot sends first: https://dedupedai.com/use-cases/rlaope-deepseek-bot/watch-a-topic-and-get-pinged/ - Have the bot read a link or image you send it: https://dedupedai.com/use-cases/rlaope-deepseek-bot/summarize-links-and-images-sent-to-bot/ ## How to add a remembering AI teammate to a Slack channel with DeepSeek Bot Repo: [rlaope/deepseek-bot](https://github.com/rlaope/deepseek-bot) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Mention the bot in Slack and ask what was decided last week; it searches its threads and answers in place. **Who it is for:** A team lead in a Slack workspace who keeps losing track of decisions made in long threads. **The problem:** Decisions get buried in old threads, and nobody wants to scroll back through weeks of messages to find them. **The outcome:** The bot answers in the thread with the earlier passage and what it implies, drawing only on the channel where you asked. ### Steps 1. Run `curl -fsSL https://raw.githubusercontent.com/rlaope/deepseek-bot/master/install.sh | bash` in a macOS terminal. 2. Run `dsh gateway setup`, choose Slack with ↑↓ and Enter, and paste the token when asked. 3. Run `dsh gateway status` to confirm the bot is up. 4. In a Slack channel, mention the bot and ask what was decided about a topic last week. ### Prompt to paste into your agent ```text Using the repo rlaope/deepseek-bot on a Mac with Node.js 22+, install it with the README's curl command, connect Slack with `dsh gateway setup`, and confirm with `dsh gateway status`. Then ask the bot in what was decided about last week; the expected output is a thread reply that quotes or points to the earlier message. Verify that the reply appears only in the channel where you asked. ``` ### Caveat The README says Slack is used live while Discord and Telegram are tested only against mocked APIs, and recall matches words, not meaning, so a paraphrase can be missed. Risk, credentials: The Slack bot token and DeepSeek API key give the bot access to your workspace and model account, so store them only where the README's setup steps put them. 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/rlaope-deepseek-bot/team-channel-memory/ ## How to get docx, xlsx or pptx reports drafted in chat with DeepSeek Bot Repo: [rlaope/deepseek-bot](https://github.com/rlaope/deepseek-bot) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Ask the bot for a report and receive the file uploaded into the conversation, ready to edit. **Who it is for:** A small business owner or analyst who needs a first draft of a report file without opening an office app. **The problem:** Starting a report from a blank file takes time, and chat messages alone are not something you can send to a client or manager. **The outcome:** The bot writes the requested docx, xlsx or pptx file and uploads it into the conversation for you to review and edit. ### Steps 1. Run `curl -fsSL https://raw.githubusercontent.com/rlaope/deepseek-bot/master/install.sh | bash` in a macOS terminal. 2. Run `dsh` to open the terminal chat with the agent. 3. Ask for the report in plain words and name the format: docx, xlsx or pptx. 4. Open the uploaded file from the conversation and check the content before you share it. ### Prompt to paste into your agent ```text Using the repo rlaope/deepseek-bot on a Mac with Node.js 22+, install it with the README's curl command and start the terminal chat with `dsh`. Ask it to draft a report on ; the expected output is a file uploaded into the conversation. Verify that the file opens and that its contents match the topic you asked for. ``` ### Caveat The README says the bot drafts the files but does not describe layout or formatting checks, so review every file before you share it. 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/rlaope-deepseek-bot/draft-report-files-in-chat/ ## How to set a watch or reminder that the bot sends first with DeepSeek Bot Repo: [rlaope/deepseek-bot](https://github.com/rlaope/deepseek-bot) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Tell the bot to watch for a change or remind you on a day, and it messages you when that happens. **Who it is for:** Someone waiting on a vendor update or a deadline who does not want to keep checking by hand. **The problem:** Checking a page or a date again and again wastes time, and the answer often arrives when you are not looking. **The outcome:** The bot checks on a schedule and messages you when the change happens, or sends the reminder on the day you named. ### Steps 1. Run `curl -fsSL https://raw.githubusercontent.com/rlaope/deepseek-bot/master/install.sh | bash` in a macOS terminal. 2. Run `dsh gateway setup` and connect the messenger you use. 3. Message the bot directly and say what to watch, for example "Tell me when this changes" followed by a link. 4. Or say "remind me on Friday" with the task, then wait for the message. ### Prompt to paste into your agent ```text Using the repo rlaope/deepseek-bot on a Mac with Node.js 22+, install it with the README's curl command and connect a messenger with `dsh gateway setup`. Send the bot a direct message saying "Tell me when this changes" with ; the expected output is a confirmation that a watch exists and a message from the bot when the page changes. Verify the watch by checking that the bot confirms it in the same thread. ``` ### Caveat The README says watches and reminders run inside the supervised launchd services, which exist only on macOS, so they stop if the services stop. 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/rlaope-deepseek-bot/watch-a-topic-and-get-pinged/ ## How to have the bot read a link or image you send it with DeepSeek Bot Repo: [rlaope/deepseek-bot](https://github.com/rlaope/deepseek-bot) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Send the bot a web page or screenshot and get a plain read of what it says, from any messenger it is connected to. **Who it is for:** A researcher or marketer who receives links and screenshots in chat and wants a quick read before deciding what to do. **The problem:** Opening every link or screenshot to get the gist takes time, and the context often sits in a chat you already have open. **The outcome:** The bot reads the page or looks at the image and replies in the conversation with what it found. ### Steps 1. Run `curl -fsSL https://raw.githubusercontent.com/rlaope/deepseek-bot/master/install.sh | bash` in a macOS terminal. 2. Run `dsh gateway setup` and connect a messenger. 3. Send the bot a direct message with a link and ask for the main points. 4. Send an image as well, if the model you use accepts images, and ask what it shows. ### Prompt to paste into your agent ```text Using the repo rlaope/deepseek-bot on a Mac with Node.js 22+, install it with the README's curl command and connect a messenger with `dsh gateway setup`. Send the bot a direct message with and ask for the main points; the expected output is a reply in the same thread summarizing the page. Verify that the summary matches the page's actual content by opening the link yourself. ``` ### Caveat The README says image reading needs a model that accepts images and that pages drawn with JavaScript are read with a headless browser, one page per run. Risk, credentials: A page that needs a sign-in is read through a session you keep for it, so use only accounts you are allowed to use and keep that session private. 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/rlaope-deepseek-bot/summarize-links-and-images-sent-to-bot/