crawd 0.8.6 → 0.9.0

package/src/backend/server.ts

@@ -1,51 +1,38 @@
 /**
- * CrawdBackend — encapsulates the Fastify+Socket.IO server, TTS, coordinator,
+ * CrawdBackend — encapsulates the Fastify+Socket.IO server, coordinator,
  * and chat system. Used by both standalone mode (backend/index.ts) and the
  * OpenClaw plugin (plugin.ts).
+ *
+ * TTS generation has been moved to the overlay (Next.js server actions).
+ * This backend now sends text-only events.
  */
 import { randomUUID } from 'crypto'
-import { writeFile, mkdir } from 'fs/promises'
-import { join } from 'path'
 import Fastify, { type FastifyInstance } from 'fastify'
-import fastifyStatic from '@fastify/static'
 import cors from '@fastify/cors'
 import { Server } from 'socket.io'
-import OpenAI from 'openai'
 import { pumpfun } from '../lib/pumpfun/v2/index.js'
 import { ChatManager } from '../lib/chat/manager.js'
 import { PumpFunChatClient } from '../lib/chat/pumpfun/client.js'
 import { YouTubeChatClient } from '../lib/chat/youtube/client.js'
-import { Coordinator, OneShotGateway, type CoordinatorConfig, type CoordinatorEvent } from './coordinator.js'
+import { Coordinator, OneShotGateway, type CoordinatorConfig, type CoordinatorEvent, type Plan, type AutonomyMode } from './coordinator.js'
 import { generateShortId } from '../lib/chat/types.js'
-import { configureTikTokTTS, generateTikTokTTS } from '../lib/tts/tiktok.js'
 import type { ChatMessage } from '../lib/chat/types.js'
 
 // ---------------------------------------------------------------------------
 // Config types
 // ---------------------------------------------------------------------------
 
-export type TtsVoiceEntry = {
-  provider: 'openai' | 'elevenlabs' | 'tiktok'
-  voice: string
-}
-
 export type CrawdConfig = {
   enabled: boolean
   port: number
   bindHost: string
-  backendUrl?: string
-  tts: {
-    chat: TtsVoiceEntry[]
-    bot: TtsVoiceEntry[]
-    openaiApiKey?: string
-    elevenlabsApiKey?: string
-    tiktokSessionId?: string
-  }
+  autonomyMode?: AutonomyMode
   vibe: {
     enabled: boolean
     intervalMs: number
     idleAfterMs: number
     sleepAfterIdleMs: number
+    batchWindowMs: number
     prompt?: string
   }
   chat: {
@@ -85,10 +72,6 @@ export class CrawdBackend {
   private config: CrawdConfig
   private logger: CrawdLogger
 
-  private openai: OpenAI | null = null
-  private elevenlabs: any = null
-  private ttsDir: string
-  private backendUrl: string
   private buildVersion: string
 
   private chatManager: ChatManager | null = null
@@ -104,17 +87,7 @@ export class CrawdBackend {
     this.config = config
     this.logger = logger ?? defaultLogger
     this.fastify = Fastify({ logger: true })
-    this.ttsDir = join(process.cwd(), 'tmp', 'tts')
-    this.backendUrl = config.backendUrl ?? `http://localhost:${config.port}`
     this.buildVersion = randomUUID()
-
-    // Initialize TTS providers based on config
-    if (config.tts.openaiApiKey) {
-      this.openai = new OpenAI({ apiKey: config.tts.openaiApiKey })
-    }
-    if (config.tts.tiktokSessionId) {
-      configureTikTokTTS(config.tts.tiktokSessionId)
-    }
   }
 
   // =========================================================================
@@ -122,23 +95,7 @@ export class CrawdBackend {
   // =========================================================================
 
   async start(): Promise<void> {
-    // Lazy-init ElevenLabs (optional dep)
-    if (this.config.tts.elevenlabsApiKey && !this.elevenlabs) {
-      try {
-        const { ElevenLabsClient } = await import('@elevenlabs/elevenlabs-js')
-        this.elevenlabs = new ElevenLabsClient({ apiKey: this.config.tts.elevenlabsApiKey })
-      } catch {
-        this.logger.warn('ElevenLabs SDK not installed, ElevenLabs TTS disabled')
-      }
-    }
-
     await this.fastify.register(cors, { origin: true })
-    await mkdir(this.ttsDir, { recursive: true })
-    await this.fastify.register(fastifyStatic, {
-      root: this.ttsDir,
-      prefix: '/tts/',
-      decorateReply: false,
-    })
 
     this.io = new Server(this.fastify.server, {
       cors: { origin: '*' },
@@ -173,7 +130,7 @@ export class CrawdBackend {
   // Public API (used by plugin tool handlers)
   // =========================================================================
 
-  /** Speak on the livestream — emits overlay event + TTS. Blocks until overlay finishes playing. */
+  /** Speak on the livestream — emits text-only overlay event. Blocks until overlay finishes. */
   async handleTalk(text: string): Promise<{ spoken: boolean }> {
     if (!text || typeof text !== 'string') {
       return { spoken: false }
@@ -182,23 +139,15 @@ export class CrawdBackend {
     this.coordinator?.notifySpeech()
 
     const id = randomUUID()
-    try {
-      const tts = await this.generateTTSWithFallback(text, this.config.tts.bot)
-      this.logger.info(`TTS generated: ${tts.url}`)
-      this.io.emit('crawd:talk', { id, message: text, ttsUrl: tts.url, ttsProvider: tts.provider })
-    } catch (e) {
-      this.logger.error('Failed to generate TTS, emitting without audio', e)
-      this.io.emit('crawd:talk', { id, message: text, ttsUrl: '' })
-    }
+    this.io.emit('crawd:talk', { id, message: text })
 
     await this.waitForAck(id)
     return { spoken: true }
   }
 
   /**
-   * Reply to a chat message — reads original aloud (chat voice),
-   * then speaks bot reply (bot voice). Emits `crawd:reply-turn`.
-   * Blocks until overlay finishes playing both audios.
+   * Reply to a chat message — emits text-only overlay event with chat + bot message.
+   * Blocks until overlay finishes.
    */
   async handleReply(
     text: string,
@@ -211,34 +160,44 @@ export class CrawdBackend {
     this.coordinator?.notifySpeech()
 
     const id = randomUUID()
-    try {
-      const [chatTts, botTts] = await Promise.all([
-        this.generateTTSWithFallback(`Chat says: ${chat.message}`, this.config.tts.chat),
-        this.generateTTSWithFallback(text, this.config.tts.bot),
-      ])
-      this.io.emit('crawd:reply-turn', {
-        id,
-        chat: { username: chat.username, message: chat.message },
-        botMessage: text,
-        chatTtsUrl: chatTts.url,
-        botTtsUrl: botTts.url,
-        chatTtsProvider: chatTts.provider,
-        botTtsProvider: botTts.provider,
-      })
-    } catch (e) {
-      this.logger.error('Failed to generate reply-turn TTS, falling back to talk', e)
-      try {
-        const tts = await this.generateTTSWithFallback(text, this.config.tts.bot)
-        this.io.emit('crawd:talk', { id, message: text, ttsUrl: tts.url, ttsProvider: tts.provider })
-      } catch {
-        this.io.emit('crawd:talk', { id, message: text, ttsUrl: '' })
-      }
-    }
+    this.io.emit('crawd:reply-turn', {
+      id,
+      chat: { username: chat.username, message: chat.message },
+      botMessage: text,
+    })
 
     await this.waitForAck(id)
     return { spoken: true }
   }
 
+  // =========================================================================
+  // Plan API (used by plugin tool handlers)
+  // =========================================================================
+
+  setPlan(goal: string, steps: string[]): { plan: Plan } | { error: string } {
+    if (!this.coordinator) return { error: 'Coordinator not enabled' }
+    const plan = this.coordinator.setPlan(goal, steps)
+    return { plan }
+  }
+
+  markPlanStepDone(step: number): { plan: Plan } | { error: string } {
+    if (!this.coordinator) return { error: 'Coordinator not enabled' }
+    const plan = this.coordinator.markStepDone(step)
+    if (!plan) return { error: 'No active plan or invalid step index' }
+    return { plan }
+  }
+
+  abandonPlan(): { plan: Plan } | { error: string } {
+    if (!this.coordinator) return { error: 'Coordinator not enabled' }
+    const plan = this.coordinator.abandonPlan()
+    if (!plan) return { error: 'No active plan to abandon' }
+    return { plan }
+  }
+
+  getPlan(): { plan: Plan | null } {
+    return { plan: this.coordinator?.getPlan() ?? null }
+  }
+
   getIO(): Server {
     return this.io
   }
@@ -265,95 +224,6 @@ export class CrawdBackend {
     }
   }
 
-  // =========================================================================
-  // TTS (with ordered fallback chain)
-  // =========================================================================
-
-  async generateTTSWithFallback(text: string, chain: TtsVoiceEntry[]): Promise<{ url: string; provider: TtsVoiceEntry['provider'] }> {
-    let lastError: Error | null = null
-
-    for (const entry of chain) {
-      try {
-        let url: string
-        switch (entry.provider) {
-          case 'elevenlabs':
-            url = await this.generateElevenLabsTTS(text, entry.voice)
-            break
-          case 'openai':
-            url = await this.generateOpenAITTS(text, entry.voice)
-            break
-          case 'tiktok':
-            url = await this.generateTikTokTTSFile(text, entry.voice)
-            break
-        }
-        return { url, provider: entry.provider }
-      } catch (e) {
-        lastError = e instanceof Error ? e : new Error(String(e))
-        this.logger.warn(`TTS ${entry.provider}/${entry.voice} failed: ${lastError.message}, trying next...`)
-      }
-    }
-
-    throw lastError ?? new Error('No TTS providers configured')
-  }
-
-  private async generateOpenAITTS(text: string, voice: string): Promise<string> {
-    if (!this.openai) throw new Error('OpenAI not configured (missing apiKey)')
-
-    const response = await this.openai.audio.speech.create({
-      model: 'gpt-4o-mini-tts',
-      voice: voice as 'onyx',
-      input: text,
-    })
-
-    const buffer = Buffer.from(await response.arrayBuffer())
-    return await this.saveTTSFile(buffer)
-  }
-
-  private async generateElevenLabsTTS(text: string, voiceId: string): Promise<string> {
-    if (!this.elevenlabs) throw new Error('ElevenLabs not configured (missing apiKey)')
-
-    const audio = await this.elevenlabs.textToSpeech.convert(voiceId, {
-      modelId: 'eleven_multilingual_v2',
-      text,
-      outputFormat: 'mp3_44100_128',
-      voiceSettings: {
-        stability: 0,
-        similarityBoost: 1.0,
-        useSpeakerBoost: true,
-        speed: 1.0,
-      },
-    })
-
-    const response = new Response(audio as any)
-    const arrayBuffer = await response.arrayBuffer()
-    const buffer = Buffer.from(arrayBuffer)
-
-    // Check if response is valid MP3
-    const isMP3 =
-      (buffer[0] === 0x49 && buffer[1] === 0x44 && buffer[2] === 0x33) ||
-      (buffer[0] === 0xff && (buffer[1] & 0xe0) === 0xe0)
-
-    if (!isMP3) {
-      const preview = buffer.subarray(0, 200).toString('utf-8')
-      throw new Error(`ElevenLabs returned non-audio response: ${preview.slice(0, 100)}`)
-    }
-
-    return await this.saveTTSFile(buffer)
-  }
-
-  private async generateTikTokTTSFile(text: string, voice?: string): Promise<string> {
-    const buffer = await generateTikTokTTS(text, voice)
-    return await this.saveTTSFile(buffer)
-  }
-
-  private async saveTTSFile(buffer: Buffer): Promise<string> {
-    const filename = `${randomUUID()}.mp3`
-    await mkdir(this.ttsDir, { recursive: true })
-    await writeFile(join(this.ttsDir, filename), buffer)
-    this.logger.info(`TTS file written: ${filename}, size: ${buffer.length} bytes`)
-    return `${this.backendUrl}/tts/${filename}`
-  }
-
   // =========================================================================
   // Chat system + Coordinator
   // =========================================================================
@@ -387,10 +257,11 @@ export class CrawdBackend {
       )
 
       const coordConfig: Partial<CoordinatorConfig> = {
-        vibeEnabled: this.config.vibe.enabled,
+        autonomyMode: this.config.autonomyMode ?? 'vibe',
         vibeIntervalMs: this.config.vibe.intervalMs,
         idleAfterMs: this.config.vibe.idleAfterMs,
         sleepAfterIdleMs: this.config.vibe.sleepAfterIdleMs,
+        batchWindowMs: this.config.vibe.batchWindowMs,
       }
       if (this.config.vibe.prompt) {
         coordConfig.vibePrompt = this.config.vibe.prompt
@@ -408,6 +279,14 @@ export class CrawdBackend {
           this.io.emit('crawd:status', { status: 'vibing' })
         } else if (event.type === 'chatProcessed') {
           this.io.emit('crawd:status', { status: 'chatting' })
+        } else if (event.type === 'planNudgeExecuted' && !event.skipped) {
+          this.io.emit('crawd:status', { status: 'planning' })
+        } else if (event.type === 'planCreated') {
+          this.io.emit('crawd:plan', { type: 'created', planId: event.planId, goal: event.goal })
+        } else if (event.type === 'planCompleted') {
+          this.io.emit('crawd:plan', { type: 'completed', planId: event.planId })
+        } else if (event.type === 'planAbandoned') {
+          this.io.emit('crawd:plan', { type: 'abandoned', planId: event.planId })
         }
       })
 
@@ -432,6 +311,11 @@ export class CrawdBackend {
         socket.emit('crawd:mcap', { mcap: this.latestMcap })
       }
 
+      // Sync current coordinator state so the overlay knows the initial animation
+      if (this.coordinator) {
+        socket.emit('crawd:status', { status: this.coordinator.state })
+      }
+
       socket.on('crawd:talk:done', (data: { id?: string }) => {
         if (data?.id) {
           this.logger.info(`Talk ack received: ${data.id}`)
@@ -495,6 +379,10 @@ export class CrawdBackend {
       return { enabled: true, ...this.coordinator.getState() }
     })
 
+    this.fastify.get('/plan', async () => {
+      return this.getPlan()
+    })
+
     this.fastify.post<{ Body: Partial<CoordinatorConfig> }>(
       '/coordinator/config',
       async (request, reply) => {
@@ -537,25 +425,14 @@ export class CrawdBackend {
           return reply.status(400).send({ error: 'username, message, and response are required' })
         }
 
-        try {
-          const [chatTts, botTts] = await Promise.all([
-            this.generateTTSWithFallback(`Chat says: ${message}`, this.config.tts.chat),
-            this.generateTTSWithFallback(response, this.config.tts.bot),
-          ])
-          this.io.emit('crawd:reply-turn', {
-            id: randomUUID(),
-            chat: { username, message },
-            botMessage: response,
-            chatTtsUrl: chatTts.url,
-            botTtsUrl: botTts.url,
-            chatTtsProvider: chatTts.provider,
-            botTtsProvider: botTts.provider,
-          })
-          return { ok: true }
-        } catch (e) {
-          this.fastify.log.error(e, 'failed to generate mock turn TTS')
-          return reply.status(500).send({ error: 'Failed to generate TTS' })
-        }
+        const id = randomUUID()
+        this.io.emit('crawd:reply-turn', {
+          id,
+          chat: { username, message },
+          botMessage: response,
+        })
+
+        return { ok: true, id }
       },
     )
   }
@@ -585,41 +462,20 @@ export class CrawdBackend {
 
 export function configFromEnv(): CrawdConfig {
   const port = Number(process.env.PORT || 4000)
-
-  const botChain: TtsVoiceEntry[] = []
-  const chatChain: TtsVoiceEntry[] = []
-
-  if (process.env.ELEVENLABS_API_KEY) {
-    botChain.push({ provider: 'elevenlabs', voice: process.env.TTS_BOT_VOICE || 'TX3LPaxmHKxFdv7VOQHJ' })
-  }
-  if (process.env.OPENAI_API_KEY) {
-    botChain.push({ provider: 'openai', voice: process.env.TTS_BOT_VOICE || 'onyx' })
-  }
-
-  if (process.env.TIKTOK_SESSION_ID) {
-    chatChain.push({ provider: 'tiktok', voice: process.env.TTS_CHAT_VOICE || 'en_us_002' })
-  }
-  if (process.env.OPENAI_API_KEY) {
-    chatChain.push({ provider: 'openai', voice: process.env.TTS_CHAT_VOICE || 'onyx' })
-  }
+  const rawMode = process.env.AUTONOMY_MODE
+  const autonomyMode = (rawMode === 'vibe' || rawMode === 'plan' || rawMode === 'none') ? rawMode : undefined
 
   return {
     enabled: true,
     port,
     bindHost: process.env.BIND_HOST || '0.0.0.0',
-    backendUrl: process.env.BACKEND_URL || `http://localhost:${port}`,
-    tts: {
-      chat: chatChain,
-      bot: botChain,
-      openaiApiKey: process.env.OPENAI_API_KEY,
-      elevenlabsApiKey: process.env.ELEVENLABS_API_KEY,
-      tiktokSessionId: process.env.TIKTOK_SESSION_ID,
-    },
+    autonomyMode,
     vibe: {
       enabled: process.env.VIBE_ENABLED !== 'false',
       intervalMs: Number(process.env.VIBE_INTERVAL_MS || 30_000),
       idleAfterMs: Number(process.env.IDLE_AFTER_MS || 180_000),
       sleepAfterIdleMs: Number(process.env.SLEEP_AFTER_IDLE_MS || 180_000),
+      batchWindowMs: Number(process.env.CHAT_BATCH_WINDOW_MS || 20_000),
       prompt: process.env.VIBE_PROMPT,
     },
     chat: {

package/src/commands/skill.ts

@@ -110,6 +110,9 @@ crawd config set vibe.idleAfter 180
 # Seconds of inactivity before going to sleep (default: 360)
 crawd config set vibe.sleepAfter 360
 
+# Chat batch throttle window in seconds (default: 20)
+crawd config set vibe.chatBatchWindow 20
+
 # Disable vibing entirely
 crawd config set vibe.enabled false
 \`\`\`

package/src/commands/start.ts

@@ -54,6 +54,7 @@ function buildEnv(config: Config): NodeJS.ProcessEnv {
   env.VIBE_INTERVAL_MS = String(config.vibe.interval * 1000)
   env.IDLE_AFTER_MS = String(config.vibe.idleAfter * 1000)
   env.SLEEP_AFTER_IDLE_MS = String((config.vibe.sleepAfter - config.vibe.idleAfter) * 1000)
+  env.CHAT_BATCH_WINDOW_MS = String(config.vibe.chatBatchWindow * 1000)
 
   env.YOUTUBE_ENABLED = String(config.chat.youtube.enabled)
   if (config.chat.youtube.videoId) {

package/src/config/schema.ts

@@ -50,6 +50,8 @@ export const ConfigSchema = z.object({
     idleAfter: z.number().default(180),
     /** Seconds of inactivity before going to sleep (must be > idleAfter) */
     sleepAfter: z.number().default(360),
+    /** Seconds for chat batch throttle window (leading-edge) */
+    chatBatchWindow: z.number().default(20),
   }).default({}),
 
   /** Stream configuration */