From 652bb6bd4c096d5c6dae5546ecac82217640860d Mon Sep 17 00:00:00 2001
From: Huaaudio <jiahua.zhang@kuleuven.be>
Date: Sat, 21 Mar 2026 07:18:14 +0100
Subject: [PATCH] update with corrent transcriber and leave_tool

---
 pkg/agent/loop.go             |  3 ++
 pkg/asr/groq_transcriber.go   | 18 ++++++++++++
 pkg/bus/types.go              |  1 +
 pkg/channels/discord/voice.go |  2 ++
 pkg/tools/voice_leave.go      | 52 +++++++++++++++++++++++++++++++++++
 pkg/voice/agent.go            |  1 +
 6 files changed, 77 insertions(+)
 create mode 100644 pkg/tools/voice_leave.go

diff --git a/pkg/agent/loop.go b/pkg/agent/loop.go
index a3a0b328d..3d1ccfa99 100644
--- a/pkg/agent/loop.go
+++ b/pkg/agent/loop.go
@@ -246,6 +246,9 @@ func registerSharedTools(
 			agent.Tools.Register(messageTool)
 		}
 
+		// Always register Voice Leave Tool (it inherently checks if channel == "discord")
+		agent.Tools.Register(tools.NewVoiceLeaveTool(msgBus))
+
 		// Send file tool (outbound media via MediaStore — store injected later by SetMediaStore)
 		if cfg.Tools.IsToolEnabled("send_file") {
 			sendFileTool := tools.NewSendFileTool(
diff --git a/pkg/asr/groq_transcriber.go b/pkg/asr/groq_transcriber.go
index a17850cb0..b0a98d858 100644
--- a/pkg/asr/groq_transcriber.go
+++ b/pkg/asr/groq_transcriber.go
@@ -10,8 +10,10 @@ import (
 	"net/http"
 	"os"
 	"path/filepath"
+	"strings"
 	"time"
 
+	"github.com/sipeed/picoclaw/pkg/config"
 	"github.com/sipeed/picoclaw/pkg/logger"
 	"github.com/sipeed/picoclaw/pkg/utils"
 )
@@ -166,3 +168,19 @@ func (t *GroqTranscriber) doRequest(ctx context.Context, requestBody *bytes.Buff
 func (t *GroqTranscriber) Name() string {
 	return "groq"
 }
+
+// DetectTranscriber inspects cfg and returns the appropriate Transcriber, or
+// nil if no supported transcription provider is configured.
+func DetectTranscriber(cfg *config.Config) Transcriber {
+	// Direct Groq provider config takes priority.
+	if key := cfg.Providers.Groq.APIKey; key != "" {
+		return NewGroqTranscriber(key)
+	}
+	// Fall back to any model-list entry that uses the groq/ protocol or is explicitly named groq.
+	for _, mc := range cfg.ModelList {
+		if (strings.HasPrefix(mc.Model, "groq/") || mc.ModelName == "groq" || mc.Model == "whisper-large-v3-turbo") && mc.APIKey != "" {
+			return NewGroqTranscriber(mc.APIKey)
+		}
+	}
+	return nil
+}
diff --git a/pkg/bus/types.go b/pkg/bus/types.go
index 15aebc345..27cf61b5f 100644
--- a/pkg/bus/types.go
+++ b/pkg/bus/types.go
@@ -70,6 +70,7 @@ type AudioChunk struct {
 // VoiceControl represents state or commands for voice sessions.
 type VoiceControl struct {
 	SessionID string `json:"session_id"`
+	ChatID    string `json:"chat_id"`
 	Type      string `json:"type"`   // "state", "command"
 	Action    string `json:"action"` // "idle", "listening", "start", "stop", "leave"
 }
diff --git a/pkg/channels/discord/voice.go b/pkg/channels/discord/voice.go
index 233cfcccb..6f9e10dfa 100644
--- a/pkg/channels/discord/voice.go
+++ b/pkg/channels/discord/voice.go
@@ -23,6 +23,7 @@ func (c *DiscordChannel) handleVoiceCommand(s *discordgo.Session, m *discordgo.M
 
 		logger.InfoCF("discord", "Joining voice channel", map[string]any{"channel": vs.ChannelID})
 		vc, err := s.ChannelVoiceJoin(c.ctx, m.GuildID, vs.ChannelID, false, false)
+		vc, err := s.ChannelVoiceJoin(c.ctx, m.GuildID, vs.ChannelID, false, false)
 		if err != nil {
 			s.ChannelMessageSend(m.ChannelID, fmt.Sprintf("Failed to join voice channel: %v", err))
 			return true
@@ -34,6 +35,7 @@ func (c *DiscordChannel) handleVoiceCommand(s *discordgo.Session, m *discordgo.M
 	} else if m.Content == "!vc leave" {
 		vc, exists := s.VoiceConnections[m.GuildID]
 		if exists && vc != nil {
+			vc.Disconnect(c.ctx)
 			vc.Disconnect(c.ctx)
 			s.ChannelMessageSend(m.ChannelID, "Left Voice Channel.")
 		} else {
diff --git a/pkg/tools/voice_leave.go b/pkg/tools/voice_leave.go
new file mode 100644
index 000000000..882e7cb9d
--- /dev/null
+++ b/pkg/tools/voice_leave.go
@@ -0,0 +1,52 @@
+package tools
+
+import (
+	"context"
+
+	"github.com/sipeed/picoclaw/pkg/bus"
+	"github.com/sipeed/picoclaw/pkg/logger"
+)
+
+type VoiceLeaveTool struct {
+	bus *bus.MessageBus
+}
+
+func NewVoiceLeaveTool(mb *bus.MessageBus) *VoiceLeaveTool {
+	return &VoiceLeaveTool{bus: mb}
+}
+
+func (t *VoiceLeaveTool) Name() string {
+	return "voice_leave"
+}
+
+func (t *VoiceLeaveTool) Description() string {
+	return "Disconnects the bot from the current voice channel. Use this tool when the user says goodbye or explicitly asks you to leave the voice chat."
+}
+
+func (t *VoiceLeaveTool) Parameters() map[string]any {
+	return map[string]any{
+		"type":       "object",
+		"properties": map[string]any{},
+	}
+}
+
+func (t *VoiceLeaveTool) Execute(ctx context.Context, args map[string]any) *ToolResult {
+	channel := ToolChannel(ctx)
+	chatID := ToolChatID(ctx)
+
+	if channel != "discord" {
+		return &ToolResult{ForLLM: "Can only leave voice channels on Discord", IsError: true}
+	}
+
+	t.bus.PublishVoiceControl(ctx, bus.VoiceControl{
+		ChatID: chatID,
+		Type:   "command",
+		Action: "leave",
+	})
+
+	logger.InfoCF("agent", "Voice command triggered via tool: leave", map[string]any{"chat_id": chatID})
+
+	return &ToolResult{
+		ForLLM: "Successfully sent disconnect command to voice adapter.",
+	}
+}
diff --git a/pkg/voice/agent.go b/pkg/voice/agent.go
index 0885e6b2e..8cd01c338 100644
--- a/pkg/voice/agent.go
+++ b/pkg/voice/agent.go
@@ -243,6 +243,7 @@ func (a *Agent) processUtterance(ctx context.Context, acc *speechAccumulator) {
 		SenderID: acc.speakerID,
 		ChatID:   acc.chatID,
 		Content:  res.Text + oralPrompt,
+		Content:  res.Text + oralPrompt,
 		Peer:     bus.Peer{Kind: "channel", ID: acc.chatID},
 		Metadata: map[string]string{
 			"is_voice": "true",