kimaki 0.4.9 → 0.4.10

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -25,6 +25,9 @@ export const KimakiDiscordPlugin = async ({ client, $, directory, }) => {
25
25
  });
26
26
  return;
27
27
  }
28
+ await client.tui.showToast({
29
+ body: { message: 'Creating Discord thread...', variant: 'info' },
30
+ });
28
31
  try {
29
32
  const result = await $ `npx -y kimaki send-to-discord ${sessionID} -d ${directory}`.text();
30
33
  const urlMatch = result.match(/https:\/\/discord\.com\/channels\/\S+/);
package/dist/voice.js CHANGED
@@ -28,7 +28,17 @@ export async function transcribeAudio({ audio, prompt, language, temperature, ge
28
28
  throw new Error('Invalid audio format');
29
29
  }
30
30
  // Build the transcription prompt
31
- let transcriptionPrompt = `Please transcribe this audio file accurately. Here is some relevant information and filenames that may be present in the audio:\n<context>\n${prompt}\n</context>\n`;
31
+ let transcriptionPrompt = `Transcribe this audio accurately. The transcription will be sent to a coding agent (like Claude Code) to execute programming tasks.
32
+
33
+ Assume the speaker is using technical and programming terminology: file paths, function names, CLI commands, package names, API names, programming concepts, etc. Prioritize technical accuracy over literal transcription - if a word sounds like a common programming term, prefer that interpretation.
34
+
35
+ If the spoken message is unclear or ambiguous, rephrase it to better convey the intended meaning for a coding agent. The goal is effective communication of the user's programming intent, not a word-for-word transcription.
36
+
37
+ Here are relevant filenames and context that may appear in the audio:
38
+ <context>
39
+ ${prompt}
40
+ </context>
41
+ `;
32
42
  if (language) {
33
43
  transcriptionPrompt += `\nThe audio is in ${language}.`;
34
44
  }
package/package.json CHANGED
@@ -2,7 +2,7 @@
2
2
  "name": "kimaki",
3
3
  "module": "index.ts",
4
4
  "type": "module",
5
- "version": "0.4.9",
5
+ "version": "0.4.10",
6
6
  "scripts": {
7
7
  "dev": "tsx --env-file .env src/cli.ts",
8
8
  "prepublishOnly": "pnpm tsc",
@@ -1,4 +1,4 @@
1
1
  ---
2
- description: Send current session to Kimaki Discord
2
+ description: Create Discord thread for current session
3
3
  ---
4
- say "Session sent to discord!"
4
+ Creating Discord thread for this session...
@@ -39,6 +39,10 @@ export const KimakiDiscordPlugin: Plugin = async ({
39
39
  return
40
40
  }
41
41
 
42
+ await client.tui.showToast({
43
+ body: { message: 'Creating Discord thread...', variant: 'info' },
44
+ })
45
+
42
46
  try {
43
47
  const result =
44
48
  await $`npx -y kimaki send-to-discord ${sessionID} -d ${directory}`.text()
package/src/voice.ts CHANGED
@@ -42,7 +42,17 @@ export async function transcribeAudio({
42
42
  }
43
43
 
44
44
  // Build the transcription prompt
45
- let transcriptionPrompt = `Please transcribe this audio file accurately. Here is some relevant information and filenames that may be present in the audio:\n<context>\n${prompt}\n</context>\n`
45
+ let transcriptionPrompt = `Transcribe this audio accurately. The transcription will be sent to a coding agent (like Claude Code) to execute programming tasks.
46
+
47
+ Assume the speaker is using technical and programming terminology: file paths, function names, CLI commands, package names, API names, programming concepts, etc. Prioritize technical accuracy over literal transcription - if a word sounds like a common programming term, prefer that interpretation.
48
+
49
+ If the spoken message is unclear or ambiguous, rephrase it to better convey the intended meaning for a coding agent. The goal is effective communication of the user's programming intent, not a word-for-word transcription.
50
+
51
+ Here are relevant filenames and context that may appear in the audio:
52
+ <context>
53
+ ${prompt}
54
+ </context>
55
+ `
46
56
  if (language) {
47
57
  transcriptionPrompt += `\nThe audio is in ${language}.`
48
58
  }