Add voice audio, capture and transcript, browser monitor, and DM UI

poc/src/puppet.ts

@@ -99,6 +99,12 @@ export interface PuppetTransport {
     reactionAdd(channelId: string, messageId: string, emoji: string): Promise<unknown>;
 }
 
+/** The audio actions a chat command can trigger. */
+export interface PuppetAudio {
+    speak(text: string, options?: { provider?: "local" | "api"; voice?: string }): Promise<unknown>;
+    stop(): void;
+}
+
 interface PollState {
     channelId: string;
     authorId: string;
@@ -142,7 +148,10 @@ export class Puppet {
     private selfId?: string;
     private queue: Promise<void> = Promise.resolve();
 
-    constructor(private readonly core: PuppetTransport) {
+    constructor(
+        private readonly core: PuppetTransport,
+        private readonly audio?: PuppetAudio
+    ) {
         try {
             const state = JSON.parse(readFileSync(STATE_FILE, "utf8")) as PersistedState;
             if (state.ownerId) this.ownerId = state.ownerId;
@@ -159,8 +168,7 @@ export class Puppet {
 
             if (state.config) {
                 this.config = {
-                    contextMessages: clamp(state.config.contextMessages ?? DEFAULT_CONFIG.contextMessages, 1, 50),
-                    contextChars: clamp(state.config.contextChars ?? DEFAULT_CONFIG.contextChars, 200, 20000),
+                    contextMessages: clamp(state.config.contextMessages ?? DEFAULT_CONFIG.contextMessages, 1, 50),                    contextChars: clamp(state.config.contextChars ?? DEFAULT_CONFIG.contextChars, 200, 20000),
                     fetchCooldownSeconds: clamp(
                         state.config.fetchCooldownSeconds ?? DEFAULT_CONFIG.fetchCooldownSeconds,
                         0,
@@ -181,6 +189,7 @@ export class Puppet {
                     }
                 };
             }
+
         } catch {
             // first run: defaults apply
         }
@@ -263,7 +272,11 @@ export class Puppet {
             writeFileSync(
                 STATE_FILE,
                 JSON.stringify(
-                    { ownerId: this.ownerId, whitelist: this.whitelist, config: this.config },
+                    {
+                        ownerId: this.ownerId,
+                        whitelist: this.whitelist,
+                        config: this.config
+                    },
                     null,
                     4
                 )
@@ -446,6 +459,8 @@ export class Puppet {
 
         if (command?.name === "whitelist") {
             await this.handleWhitelistCommand(command, author.id, tier, channelId);
+        } else if (command?.name === "speak") {
+            await this.handleSpeakCommand(command, author.id, channelId);
         } else if (POLL_ENABLED && command?.name === "poll") {
             await this.handlePollCommand(author.id, channelId, text);
         }
@@ -541,6 +556,51 @@ export class Puppet {
         }
     }
 
+    /* Text to speech */
+
+    private async handleSpeakCommand(command: ParsedCommand, authorId: string, channelId: string) {
+        if (!this.audio) {
+            return this.respond(authorId, channelId, "speak", "Voice audio is not available.");
+        }
+
+        const text = command.args.join(" ").trim();
+        if (!text || text === "help") {
+            return this.respond(
+                authorId,
+                channelId,
+                "speak",
+                ["speak — text to speech.", "Usage:", "  speak <text>", "  speak stop"].join("\n")
+            );
+        }
+
+        if (text === "stop") {
+            this.audio.stop();
+            this.actionLog.append({ type: "command", command: "speak stop", requesterId: authorId, result: "ok" });
+            return this.respond(authorId, channelId, "speak", "Stopped.");
+        }
+
+        try {
+            await this.audio.speak(text, {});
+            this.actionLog.append({
+                type: "command",
+                command: "speak",
+                requesterId: authorId,
+                result: "queued",
+                chars: text.length
+            });
+            return this.respond(authorId, channelId, "speak", "Speaking.");
+        } catch (error) {
+            this.actionLog.append({
+                type: "command",
+                command: "speak",
+                requesterId: authorId,
+                result: "error",
+                error: String(error)
+            });
+            return this.respond(authorId, channelId, "speak", `Could not speak: ${(error as Error).message}`);
+        }
+    }
+
     /* Polls */
 
     private async handlePollCommand(authorId: string, channelId: string, text: string) {