whisper-windows-mcp 2.0.0 → 2.2.0

package/.github/ISSUE_TEMPLATE/bug_report.md

@@ -0,0 +1,42 @@
+---
+name: Bug report
+about: Something isn't working correctly
+title: '[BUG] '
+labels: bug
+assignees: ''
+---
+
+## What happened
+
+<!-- Describe what went wrong -->
+
+## Steps to reproduce
+
+1. 
+2. 
+3. 
+
+## Expected behavior
+
+<!-- What should have happened instead -->
+
+## Environment
+
+- **OS:** Windows [version — run `winver`]
+- **Node.js:** [run `node --version`]
+- **whisper-windows-mcp version:** [shown in Claude Desktop Settings → Developer]
+- **GPU:** [model name]
+- **Vulkan build:** Yes / No
+- **Model:** [e.g. ggml-medium.en.bin, ggml-large-v3.bin]
+
+## Error output
+
+<!-- Paste any error messages here -->
+
+```
+error output here
+```
+
+## Additional context
+
+<!-- Anything else that might help — file format, file duration, language, etc. -->

package/.github/ISSUE_TEMPLATE/feature_request.md

@@ -0,0 +1,29 @@
+---
+name: Feature request
+about: Suggest a new feature or improvement
+title: '[FEATURE] '
+labels: enhancement
+assignees: ''
+---
+
+## What problem would this solve
+
+<!-- Describe the use case. What are you trying to do that isn't possible or is difficult right now? -->
+
+## Proposed solution
+
+<!-- How would you like this to work? Be as specific as you can. -->
+
+## Alternatives considered
+
+<!-- Have you tried other approaches? What didn't work? -->
+
+## Hardware context
+
+<!-- Relevant if this is GPU/model related -->
+- GPU: 
+- Model size you're using:
+
+## Additional context
+
+<!-- Anything else — links to similar implementations, whisper.cpp flags that might support this, etc. -->

package/.github/ISSUE_TEMPLATE/gpu_test_result.md

@@ -0,0 +1,32 @@
+---
+name: GPU test result
+about: Share your GPU acceleration test results to help other users
+title: '[GPU] '
+labels: hardware-report
+assignees: ''
+---
+
+## Hardware
+
+- **GPU:** 
+- **VRAM:** 
+- **CPU:** 
+- **RAM:** 
+
+## Test
+
+- **Model used:** [e.g. ggml-medium.en.bin, ggml-large-v3.bin]
+- **File duration:** [e.g. 5 minutes]
+- **Language:** 
+
+## Results
+
+- **GPU utilization during transcription:** [e.g. 20%]
+- **Time to complete:** 
+- **CPU utilization during transcription:** 
+- **Vulkan build:** Yes / No
+- **Any issues:** 
+
+## Notes
+
+<!-- Anything else worth knowing for users with similar hardware -->

package/.github/pull_request_template.md

@@ -0,0 +1,28 @@
+## What this PR does
+
+<!-- Brief description of the change -->
+
+## Type of change
+
+- [ ] Bug fix
+- [ ] New feature
+- [ ] Documentation update
+- [ ] Performance improvement
+- [ ] Other
+
+## Testing
+
+- [ ] Tested on Windows
+- [ ] Built successfully (`npm run build`)
+- [ ] Tested in Claude Desktop with a real audio file
+
+## Checklist
+
+- [ ] Version bumped in `package.json` and `src/index.ts`
+- [ ] README updated if new tools or parameters added
+- [ ] TROUBLESHOOTING.md updated if fixing a known issue
+- [ ] No new external runtime dependencies added
+
+## Related issues
+
+<!-- Closes #xxx -->

package/CODE_OF_CONDUCT.md

@@ -0,0 +1,30 @@
+# Code of Conduct
+
+## Our standards
+
+This project is built to be useful to a wide range of people — developers, journalists, legal professionals, researchers, language students, video editors, and anyone else who needs local audio transcription on Windows.
+
+We expect contributors and community members to:
+
+- Be respectful and constructive in issues and pull requests
+- Focus feedback on the work, not the person
+- Accept that different people have different use cases and hardware setups
+- Help newcomers get set up — not everyone is a developer
+
+## Unacceptable behavior
+
+- Harassment, personal attacks, or discriminatory language
+- Dismissing bug reports because they involve non-English languages or non-NVIDIA hardware
+- Spam or off-topic promotion
+
+## Enforcement
+
+Issues or pull requests that violate these standards will be closed. Repeat violations may result in being blocked from the repository.
+
+## Reporting
+
+If you experience or witness unacceptable behavior, open a private vulnerability report or contact the maintainer directly through GitHub.
+
+---
+
+This Code of Conduct is intentionally minimal. The goal is a functional, welcoming project — not bureaucracy.

package/CONTRIBUTING.md

@@ -0,0 +1,89 @@
+# Contributing to whisper-windows-mcp
+
+Thanks for your interest in contributing. This is a Windows-native MCP server for local audio transcription — contributions that improve Windows compatibility, GPU support, and usability for non-technical users are especially welcome.
+
+---
+
+## Before you start
+
+Check the [open issues](https://github.com/eviscerations/whisper-windows-mcp/issues) before starting work to avoid duplicating effort. If you want to work on something not listed, open an issue first to discuss it.
+
+See [ROADMAP.md](ROADMAP.md) for planned features and known bugs.
+
+---
+
+## What we need most
+
+- **GPU acceleration testing** — if you've tested on NVIDIA, Intel Arc, or AMD hardware not listed in the README, please share your results (GPU model, VRAM, model size, observed throughput)
+- **Bug reports with reproduction steps** — especially for Windows-specific issues
+- **Multilingual testing** — results with non-English audio and the large-v3 model
+- **Documentation improvements** — especially for non-English speakers
+
+---
+
+## Development setup
+
+```
+git clone https://github.com/eviscerations/whisper-windows-mcp
+cd whisper-windows-mcp
+npm install
+npm run build
+```
+
+The server is a single TypeScript file at `src/index.ts`. The compiled output goes to `dist/`.
+
+To test locally, update your `claude_desktop_config.json` to point at the local build:
+
+```json
+{
+  "mcpServers": {
+    "whisper": {
+      "command": "node",
+      "args": ["C:\\path\\to\\whisper-windows-mcp\\dist\\index.js"],
+      "env": {
+        "WHISPER_CLI_PATH": "C:\\whisper\\Release\\whisper-cli.exe",
+        "WHISPER_MODEL": "C:\\whisper\\models\\ggml-medium.en.bin"
+      }
+    }
+  }
+}
+```
+
+Restart Claude Desktop after each build to pick up changes.
+
+---
+
+## Pull request guidelines
+
+- Keep PRs focused — one fix or feature per PR
+- Test on Windows (this is a Windows-only tool)
+- If adding a new tool, update the tool list in README.md
+- If fixing a bug, add a note to TROUBLESHOOTING.md if it's something other users will hit
+- Bump the version in `package.json` and the version string in `src/index.ts` before submitting
+
+---
+
+## Code style
+
+- TypeScript, no external runtime dependencies beyond `@modelcontextprotocol/sdk`
+- All file I/O must handle non-ASCII filenames (Unicode, Japanese, Chinese, emoji)
+- New tools go in the `CallToolRequestSchema` handler following the existing pattern
+- Error messages should be actionable — tell the user what to do, not just what went wrong
+
+---
+
+## Reporting bugs
+
+Open an issue with:
+- Your OS version (`winver`)
+- Node.js version (`node --version`)
+- whisper-windows-mcp version
+- GPU model and whether you're using the Vulkan build
+- Exact steps to reproduce
+- Full error output if available
+
+---
+
+## License
+
+By contributing, you agree that your contributions will be licensed under the MIT license.