# AI use cases for Kaas (bybit-exchange/kaas) Repo: [bybit-exchange/kaas](https://github.com/bybit-exchange/kaas) | Use cases: 4 | Page: https://dedupedai.com/use-cases/bybit-exchange-kaas/ KaaS is a self-hosted tool that compiles notes, documents and transcripts into a searchable Markdown wiki using LLM phases, and serves it to chat and MCP clients without embeddings. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Compile your notes into a searchable Markdown wiki: https://dedupedai.com/use-cases/bybit-exchange-kaas/compile-personal-notes-into-wiki/ - Let Claude Code query your compiled wiki locally: https://dedupedai.com/use-cases/bybit-exchange-kaas/give-claude-code-access-to-wiki/ - Share the wiki with remote agents over MCP: https://dedupedai.com/use-cases/bybit-exchange-kaas/share-wiki-with-remote-agents-over-mcp/ - Keep team knowledge when people change roles: https://dedupedai.com/use-cases/bybit-exchange-kaas/keep-team-knowledge-when-people-change-roles/ ## How to compile your notes into a searchable Markdown wiki with Kaas Repo: [bybit-exchange/kaas](https://github.com/bybit-exchange/kaas) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Turn scattered notes, documents and transcripts into browsable Markdown articles you can query. **Who it is for:** A knowledge worker with notes, documents and meeting transcripts spread across folders who keeps re-answering the same questions. **The problem:** Raw notes are hard to search and chunk-based retrieval gives fragments without context. **The outcome:** A tree of Markdown articles on disk, a web UI to browse and chat, and chat answers that cite the articles behind them. ### Steps 1. Install KaaS with `curl -fsSL https://raw.githubusercontent.com/bybit-exchange/kaas/main/install.sh | sh` or use the Docker option from the Quick Start. 2. Set `LLM_API_KEY`, `LLM_BASE_URL` and `LLM_MODEL` to your OpenAI-compatible endpoint. 3. Start the service with `kaas serve` (CLI) or the Docker run command from the Quick Start. 4. Open http://localhost:8080, submit your text, file or URL on the Submit page, and wait for the compile to finish. 5. Browse the Wiki page and ask questions in Chat, then open the cited articles to check them. ### Prompt to paste into your agent ```text Using the KaaS repo at bybit-exchange/kaas, set up the service and compile the files in into the wiki. Expected output: a list of Markdown articles under the data directory and one chat answer that cites at least one article. Verify that the cited article exists on disk. ``` ### Caveat Compile quality and speed depend on the LLM you configure, and the README notes slow local models may need longer write and extract timeouts. Risk, credentials: Your LLM API key is passed as an environment variable and your source documents are sent to the configured LLM endpoint. 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/bybit-exchange-kaas/compile-personal-notes-into-wiki/ ## How to let Claude Code query your compiled wiki locally with Kaas Repo: [bybit-exchange/kaas](https://github.com/bybit-exchange/kaas) | Difficulty: Advanced | Time: 45 minutes | Basis: From the README Add the kb-ai stdio MCP server so Claude Code can ask the wiki questions with cited answers. **Who it is for:** A developer working in Claude Code who wants project knowledge from compiled notes available without leaving the editor. **The problem:** Coding agents have no access to the decisions and context stored in your personal or team notes. **The outcome:** Claude Code gets a single ask tool that returns a cited Markdown answer grounded in the wiki. ### Steps 1. Build or install a knowledge base as described in the Quick Start so the wiki exists on disk. 2. Set `KAAS_KB_DIR` to the knowledge-base root and the `LLM_*` credentials in your environment. 3. Register the server with `claude mcp add kaas -- kb-ai mcp`. 4. Ask Claude Code a question that the wiki should answer and check that the reply cites wiki articles. ### Prompt to paste into your agent ```text Using the KaaS repo at bybit-exchange/kaas, register the stdio MCP server for the knowledge base at with Claude Code. Expected output: a Claude Code session where the ask tool returns a cited Markdown answer. Verify that the cited article path exists under . ``` ### Caveat The README does not say how to install the kb-ai command on its own, so confirm it is on your PATH before registering it. Risk, credentials: The stdio server reads LLM_* credentials from the environment that the agent launches it with. 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/bybit-exchange-kaas/give-claude-code-access-to-wiki/ ## How to share the wiki with remote agents over MCP with Kaas Repo: [bybit-exchange/kaas](https://github.com/bybit-exchange/kaas) | Difficulty: Advanced | Time: 60 minutes | Basis: From the README Run KaaS with remote MCP enabled and a bearer token so other machines can query it. **Who it is for:** A team lead running a shared KaaS server on an intranet, where several engineers' agents need the same knowledge. **The problem:** Each engineer's agent would otherwise need its own local copy and setup of the knowledge base. **The outcome:** Remote MCP clients on the network call the ask tool at a single /mcp endpoint, authenticated with a token. ### Steps 1. Start the container with `KAAS_MCP_ENABLED=true` and `KAAS_MCP_TOKEN` set to a long random secret, using the Docker command in the Quick Start. 2. Confirm the MCP endpoint is reachable at http://host:8080/mcp from a client machine. 3. On each client, run `claude mcp add --transport http kaas http://host:8080/mcp` and set the Authorization header to the bearer token. 4. Ask a question from a client and confirm it returns a cited answer. ### Prompt to paste into your agent ```text Using the KaaS repo at bybit-exchange/kaas, enable remote MCP on the server at with a bearer token and connect one client. Expected output: a successful ask call from the remote client returning a cited answer. Verify that a request without the token is rejected. ``` ### Caveat The README says token auth is off by default and assumes a local or intranet deployment, so do not expose the endpoint to the public internet without extra protection. Risk, credentials: Bearer token auth is off by default, so you must set KAAS_MCP_TOKEN or anyone who can reach port 8080 can query the wiki. 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/bybit-exchange-kaas/share-wiki-with-remote-agents-over-mcp/ ## How to keep team knowledge when people change roles with Kaas Repo: [bybit-exchange/kaas](https://github.com/bybit-exchange/kaas) | Difficulty: Intermediate | Time: 60 minutes | Basis: From the README Compile each role's material into a wiki so the context survives staff turnover. **Who it is for:** An operations or engineering lead whose team loses context every time someone changes roles or leaves. **The problem:** Context built up by individuals walks out the door with them and new hires spend weeks rebuilding it. **The outcome:** A role-oriented wiki of distilled articles that stays with the seat when the person in it changes. ### Steps 1. Collect the role's documents, meeting transcripts and notes into one folder. 2. Install KaaS and set your LLM credentials as described in the Quick Start. 3. Submit the folder's files through the Submit page or start the service with `kaas serve`. 4. Review the compiled articles in the Wiki page and hand-edit any article that needs correction. ### Prompt to paste into your agent ```text Using the KaaS repo at bybit-exchange/kaas, compile the documents for the seat from into a wiki. Expected output: Markdown articles for that role that answer the team's common questions. Verify that a sample answer cites an article containing the source facts. ``` ### Caveat The README describes the role-based approach as its own motivation, and output quality depends on the source material and chosen LLM. Risk, credentials: Role documents and transcripts are sent to the configured LLM endpoint, so confirm your provider meets your data policy before compiling sensitive material. 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/bybit-exchange-kaas/keep-team-knowledge-when-people-change-roles/