Type4Me is a macOS voice-input application that transcribes speech, improves or translates the resulting text with language models, and stores data locally. Mac users use it for dictation, text editing, translation, questions about selected text, and voice-controlled system actions; its catalogue instructions help agents manage its custom vocabulary.
Borrowing it
Nothing to install: this file belongs to joewongjc/type4me. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/joewongjc/type4me/main/AGENTS.mdgit clone --depth 1 https://github.com/joewongjc/type4meWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/joewongjc/type4me/agents-md)<a href="https://agentmods.dev/instructions/joewongjc/type4me/agents-md"><img src="https://agentmods.dev/badge/instructions/joewongjc/type4me/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/joewongjc/type4me/agents-md"><img src="https://agentmods.dev/badge/instructions/joewongjc/type4me/agents-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.06394 | $0.06394 |
| Opus 5.5 | $0.02558 | $0.02558 |
| Sonnet 5.5 | $0.01279 | $0.01279 |
| Haiku 4.5 | $0.00639 | $0.00639 |
Grade C, and why
type4me AGENTS.md scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
**Important**: SPM caches manifest evaluation in `~/Library/Caches/org.swift.swiftpm`. When switching variants manually (not via build-dmg.sh), clear this cache: `rm -rf .build ~/Library/Caches/org.swift.swiftpm` How it starts
The opening of the file, as written. The whole thing — 358 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Type4Me — Development Guide
Overview
macOS menu bar voice input tool with dual-engine local ASR, multi-provider cloud ASR, and LLM post-processing.
Local ASR: SenseVoice via native sherpa-onnx (streaming) + Qwen3-ASR (final calibration, Python WebSocket service managed by SenseVoiceServerManager).
Cloud ASR: 18 providers implemented (Volcano, StepFun, MiMo batch, OpenAI, Microsoft Azure MAI, Deepgram, Cartesia, AssemblyAI, ElevenLabs, Gemini, Grok, Soniox, Bailian, Baidu, Meta Muse, Groq batch, Mistral AI batch, Tencent Cloud), plus Apple Speech. StepFun, Bailian, Gemini, ElevenLabs, Grok, Microsoft Azure, and Tencent Cloud let users switch between real-time streaming and batch recognition.
Swift Package Manager project, no Xcode project file. Optional sherpa-onnx.xcframework enables local SenseVoice, Silero VAD, and punctuation restoration.
Machine-Local Agent Notes
A maintainer machine may keep uncommitted notes at .agents/LOCAL.md in the main checkout (machine paths, local-only assets, release skills under .agents/skills/). Before working, read it if it exists; from a linked worktree, find it next to the main checkout via git rev-parse --path-format=absolute --git-common-dir. Never commit .agents/LOCAL.md, local skills, or key files such as *.p8.
Branch Naming and Lifecycle
Use one short-lived branch for one independently reviewable change. Branches must use lowercase ASCII kebab-case and follow this form:
<type>/<issue-number-optional>-<concise-description>
Examples:
feat/241-cartesia-streaming-asr
fix/247-agent-context-cleanup
docs/intelli-sense-pangu-terminology
perf/reduce-startup-latency
refactor/session-output-pipeline
test/recognition-session-recovery
chore/update-swift-format
release/2.1.1
hotfix/251-crash-on-launch
Allowed types
| Type | Use for |
|---|---|
feat |
A user-visible capability or substantial new behavior. |
fix |
A defect fix. Prefix the issue number when one exists. |
docs |
Documentation-only changes. |
perf |
A measurable performance improvement without a primary behavior change. |
refactor |
Internal restructuring with no intended behavior change. |
test |
Test-only additions or repairs. |
chore |
Tooling, repository maintenance, or non-product housekeeping. |
release |
Release preparation and versioning. |
hotfix |
An urgent production fix. |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago Changed · +8 lines · +262 tokens per session 763be980783b
- 11d ago Changed · +118 tokens per session 1e7b14f2ea4a
- 12d ago Changed · +1 lines · +87 tokens per session d208154f1f26
- 17d ago Changed · +7 lines · +219 tokens per session 360bc23a8f71
- 21d ago Changed · +10 tokens per session 2c86802f7a5e
- 22d ago Changed · +4 lines · +361 tokens per session aae60e255592
- 24d ago Changed · +2 lines · +96 tokens per session e5d464a4f885
- 26d ago Changed · +3 lines · +266 tokens per session 9315193da499
- 1mo ago Changed · +5 lines · +139 tokens per session e6583a1c04f6
- 1mo ago First seen · 328 lines · 4,836 tokens per session scan C 7950f495e6a6
type4me AGENTS.md is an instructions file published in the GitHub repository joewongjc/type4me (1,541 stars, last pushed 6d ago), licensed MIT. It adds 6,394 tokens to every session, about $0.0256 per session on Opus 5.5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
type4me AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| type4me AGENTS.mdjoewongjc/type4me | Instructions file | 1.5k | 6d ago | MIT | C | 6,394 | Codex, OpenCode |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | yesterday | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | yesterday | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | yesterday | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |