đ Project Ideas
Generating project ideas…
Summary
- Automatically detects and removes unused code while respecting project conventions (generated files, test fixtures, indirect references).
- Core value: Oneâcommand deadâcode elimination that matches what senior engineers would manually specify.
Details
| Key |
Value |
| Target Audience |
Developers maintaining large codebases |
| Core Feature |
LLMâpowered static analysis that infers exclusion rules from repo structure and applies them |
| Tech Stack |
Python, Treeâsitter, LLM API (OpenAI or local), CLI, optional VSâŻCode extension |
| Difficulty |
Medium |
| Monetization |
Hobby |
Notes
- Addresses akerstenâs frustration: âIf I '/goal remove unused code' in Claude today, I would not even think to specify 'check for indirect references' and 'don't consider generated code dead code'âŚâ. This tool builds those intuitions in.
- Provides concrete utility for code hygiene and could be showcased as a Show HN post, sparking discussion on AIâassisted refactoring.
Summary
- A lowâcode orchestrator that turns a naturalâlanguage goal into a coordinated team of LLM agents, each with a specific role (planner, coder, reviewer, tester) and access to dev tools (git, linters, build).
- Core value: Eliminates the need to manually prompt each step; users get a working solution with minimal oversight.
Details
| Key |
Value |
| Target Audience |
Product managers, indie hackers, teams wanting to prototype features via AI |
| Core Feature |
Goalâdriven workflow engine that spawns, communicates, and aggregates results from multiple agents |
| Tech Stack |
Node.js/TypeScript, LangChain/LlamaIndex, Redis for message passing, Docker sandboxed agents, OpenAI API |
| Difficulty |
High |
| Monetization |
Revenue-ready: subscription tiered by agentâhours |
Notes
- Directly answers fakeBeerDrinkerâs joke about an âagentic mapreduceâ and a âtalent denseâ team of prompters, delivering a realizable version.
- Could ignite HN debate on agent safety, cost control, and the future of AIâaugmented software teams.
Summary
- A curated marketplace of expertâcrafted, battleâtested prompt templates for common developer chores (deadâcode detection, security audit, performance profiling, etc.) that include the necessary nuance and tool usage.
- Core value: Saves time and improves reliability by reusing prompts that already encode indirectâreference checks, file exclusions, etc.
Details
| Key |
Value |
| Target Audience |
Developers who use LLMs for coding assistance, prompt engineers looking to monetize expertise |
| Core Feature |
Searchable repository, versioning, ability to run prompts via CLI or IDE plugin with variable substitution |
| Tech Stack |
React frontend, PostgreSQL DB, Node.js backend, OAuth, Stripe for payments |
| Difficulty |
LowâMedium |
| Monetization |
Revenue-ready: perâprompt purchase or subscription |
Notes
- Mirrors akerstenâs desire to avoid repeatedly specifying details like âexclude generated files and test fixturesâ; the library supplies readyâmade prompts that already contain them.
- Offers a practical venue for HN users to share and discover highâquality prompts, encouraging community discussion and reuse.