# AI use cases for Nexus Pilot (GuqierMcl/NexusPilot) Repo: [GuqierMcl/NexusPilot](https://github.com/GuqierMcl/NexusPilot) | Use cases: 4 | Page: https://dedupedai.com/use-cases/guqiermcl-nexuspilot/ NexusPilot is an open-source desktop workbench that lets you query and manage PostgreSQL, MySQL, SQLite, Redis, ClickHouse, and Oracle databases through a natural-language AI agent. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Ask questions about a database in plain language: https://dedupedai.com/use-cases/guqiermcl-nexuspilot/ask-questions-about-a-database-in-plain-language/ - Run a multi-step data task with approval checks: https://dedupedai.com/use-cases/guqiermcl-nexuspilot/run-a-multi-step-data-task-with-approval/ - Inspect ClickHouse tables, projections, and indexes: https://dedupedai.com/use-cases/guqiermcl-nexuspilot/inspect-clickhouse-objects-and-indexes/ - Verify a Nexus Pilot change before a pull request: https://dedupedai.com/use-cases/guqiermcl-nexuspilot/verify-a-nexuspilot-change-before-a-pull-request/ ## How to ask questions about a database in plain language with Nexus Pilot Repo: [GuqierMcl/NexusPilot](https://github.com/GuqierMcl/NexusPilot) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Connect a database and get answers from its data without writing SQL by hand. **Who it is for:** An analyst or developer who needs a quick answer from a live PostgreSQL, MySQL, SQLite, or ClickHouse database and does not want to write every query manually. **The problem:** Writing and checking queries by hand for each new data question is slow, especially when the schema is unfamiliar. **The outcome:** You get query results and explanations for plain-language questions, while the agent works through the connection you set up in the workbench. ### Steps 1. Download the installer for your platform from the official release page at https://nexuspilot.dev/releases. 2. Open Nexus Pilot and add a connection to one of the supported databases: PostgreSQL, MySQL, SQLite, Redis, ClickHouse, or Oracle. 3. Open the local AI Runtime, add a provider and model, and enter the provider credentials in the app's provider settings. 4. Open the agent in Ask or Query mode and type a data question in natural language. 5. Read the query the agent ran and the returned rows before you rely on the answer. ### Prompt to paste into your agent ```text Using the NexusPilot repo at https://github.com/GuqierMcl/NexusPilot, set up a connection to and ask the agent in Ask mode: . Expected output: a written answer with the query used and the result rows. Verify that the query only reads data and that the answer matches a manual check of at least one row. ``` ### Caveat The README states that exact capabilities depend on the engine, server version, connection permissions, and runtime capability model, so some questions may not be supported for every source. Risk, credentials: Database connection details and LLM provider credentials are sensitive, so use a read-only database user where possible and do not paste credentials into shared chats or screenshots. 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/guqiermcl-nexuspilot/ask-questions-about-a-database-in-plain-language/ ## How to run a multi-step data task with approval checks with Nexus Pilot Repo: [GuqierMcl/NexusPilot](https://github.com/GuqierMcl/NexusPilot) | Difficulty: Advanced | Time: 90 minutes | Basis: From the README Delegate a multi-step database task to the Agent mode and review each write before it runs. **Who it is for:** A data engineer or developer who needs the agent to explore a database and make a sequence of changes, and who wants to review risky writes first. **The problem:** Multi-step database tasks are error-prone when done by hand, and an automated agent could make accidental writes without a clear review step. **The outcome:** The agent carries out the task through restricted workbench tools, and SQL or Redis writes stop for risk analysis and approval. ### Steps 1. Install Bun 1.3 or later and a Rust toolchain, then run `bun install` in the cloned repository. 2. Run `bun run dev:all` to start the desktop app and the AI Runtime together. 3. Add a SQL or Redis connection in the app and open the agent in Agent mode. 4. Describe the multi-step task in natural language and let the agent explore the objects it needs. 5. Review the risk analysis and approval prompt for each SQL or Redis write, and confirm only the changes you expect. ### Prompt to paste into your agent ```text Using the NexusPilot repo at https://github.com/GuqierMcl/NexusPilot, run `bun install` and `bun run dev:all`, then give the Agent mode this task on : . Expected output: a summary of each step and each approved write. Verify that every write was shown with a risk analysis and approved by a person before it ran. ``` ### Caveat The README says SQL and Redis writes currently use risk analysis and approval flows, so other write paths may not have the same review step. Risk, credentials: Agent writes run against the connected database with the permissions of the connection, so use an account with only the rights the task needs. 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/guqiermcl-nexuspilot/run-a-multi-step-data-task-with-approval/ ## How to inspect ClickHouse tables, projections, and indexes with Nexus Pilot Repo: [GuqierMcl/NexusPilot](https://github.com/GuqierMcl/NexusPilot) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Browse ClickHouse databases, views, projections, and data-skipping indexes in one tree. **Who it is for:** A data team member who manages ClickHouse analytics tables and needs to check how objects and indexes are set up. **The problem:** ClickHouse objects such as projections and data-skipping indexes are hard to review when you only have a generic table view. **The outcome:** You see the ClickHouse databases, tables, views, projections, and data-skipping indexes in the workbench's native object tree and can ask the agent about them. ### Steps 1. Download the installer for your platform from https://nexuspilot.dev/releases and open Nexus Pilot. 2. Add a ClickHouse connection in the connection tree. 3. Expand the connection to open its databases and tables. 4. Open the views, projections, and data-skipping indexes for a table to review their structure. 5. Ask the agent in Ask mode to explain how a given table's projection or index is defined. ### Prompt to paste into your agent ```text Using the NexusPilot repo at https://github.com/GuqierMcl/NexusPilot, connect to a ClickHouse server at and list the projections and data-skipping indexes on table . Expected output: a list of each projection and index with its definition. Verify the list against the ClickHouse server's own metadata for that table. ``` ### Caveat The README describes native ClickHouse workflows for these objects, but it does not say every management action is available, so check what the tree allows on your server. 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/guqiermcl-nexuspilot/inspect-clickhouse-objects-and-indexes/ ## How to verify a Nexus Pilot change before a pull request with Nexus Pilot Repo: [GuqierMcl/NexusPilot](https://github.com/GuqierMcl/NexusPilot) | Difficulty: Advanced | Time: 90 minutes | Basis: From the README Run the project's type, test, format, and lint checks on a change before submitting it. **Who it is for:** A contributor or coding agent who has changed the Nexus Pilot frontend, the AI Runtime, or the Rust backend and needs to meet the project's checks before opening a pull request. **The problem:** Changes that skip the project's checks are likely to fail review, and the checks cover three separate toolchains. **The outcome:** You confirm the frontend types, AI Runtime tests, Rust tests, formatting, and clippy all pass before submitting the change. ### Steps 1. Install Bun 1.3 or later and a Rust toolchain, then run `bun install` in the repository. 2. Run `bun run tsc --noEmit` for the frontend type-check and `bun run ai-runtime:test` for the AI Runtime tests. 3. Run `cargo test --manifest-path src-tauri/Cargo.toml` and `cargo fmt --manifest-path src-tauri/Cargo.toml -- --check`. 4. Run `cargo clippy --manifest-path src-tauri/Cargo.toml -- -D warnings` and fix any warnings. 5. Read AGENTS.md for the coding guidelines, and avoid including credentials, private connection strings, or generated artifacts in the pull request. ### Prompt to paste into your agent ```text Using the NexusPilot repo at https://github.com/GuqierMcl/NexusPilot, run the verification commands in the README on my current change and report the result of each one. Expected output: a pass or fail line for tsc, ai-runtime tests, cargo test, cargo fmt check, and cargo clippy. Verify that no command was skipped and that any failure includes the first error message. ``` ### Caveat The README asks contributors to read AGENTS.md before making changes, and those guidelines may add requirements beyond the listed commands. 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/guqiermcl-nexuspilot/verify-a-nexuspilot-change-before-a-pull-request/