#Hi

1 messages · Page 1 of 1 (latest)

tawny laurelBOT
#
AI-generated Answer for .mooserini

Hi! How can I help you with ElevenLabs products or API today? Tell me what you’re building or stuck on, and I’ll pull the exact steps from the official docs.

forest cloud
#

Oh sweeet. Guess I didn’t notice you replied. That’s on me. I’ve had a lot going on lately. My goal is super simple. I rub my ai locally on my own computer with Ollama. So. The logic part is a lot easier than it would be for most. I take care of the compute power and central location for all the big work to be done. I just want to give my discord bot that I have here an upgrade from gTTS without pretending I’m some rich developer. It’s like yeah my idea is cool but it’s not unique anymore so. Doubt it’s going to be a special that people pay me to talk to

tawny laurelBOT
#
AI-generated Answer for .mooserini

Got it. If you want to replace gTTS in your Discord bot without big spend, here’s the lean, documented path:

  • Pick a low-cost TTS model

    • Use Flash v2.5 (ultra‑low latency ~75 ms) or Turbo v2.5. These are billed at $0.05 per 1,000 characters on all plans, including Free/Pay‑as‑you‑go. (elevenlabs.io)

  • Get an API key and cap spend

    • Create an API key and (optionally) restrict scopes and set a per‑key credit limit. Authenticate using the xi-api-key header. (elevenlabs.io)
    • The API is included in all plans; you only pay for generation. (help.elevenlabs.io)

  • Choose a voice

    • You can use default voices or your own; fetch voice_id via the List voices API or copy it from the UI. Note: Voice Library access via API is not available on the free tier. (elevenlabs.io)

  • Generate audio for Discord playback

    • HTTP (single shot or streamed):
      • Create speech (REST): POST /v1/text-to-speech/:voice_id with text, model_id, and optional output_format (e.g., mp3_44100_128). (elevenlabs.io)
      • Stream speech (low‑latency chunked HTTP): POST /v1/text-to-speech/:voice_id/stream. Official SDKs show how to consume the stream. (elevenlabs.io)
    • WebSocket (real‑time, word alignment available):
      • wss://api.elevenlabs.io/v1/text-to-speech/:voice_id/stream-input supports sending partial text and receiving audio chunks in real time. (elevenlabs.io)
    • Pick an output format that fits Discord’s voice pipeline:
      • Supported API formats include MP3 (22.05–44.1 kHz), PCM (up to 48 kHz), Opus (48 kHz), and μ‑law/A‑law (8 kHz). Select an output_format accordingly (e.g., MP3 defaults, or a PCM/Opus 48 kHz option if your bot can play/forward it directly). (elevenlabs.io)

  • Control delivery and speed

    • You can adjust speaking speed via the Speed/voice settings parameter; recommended range is 0.7–1.2. (help.elevenlabs.io)

  • Know limits and error handling

    • Character limits per request (useful if you buffer Discord messages): Flash v2.5/Turbo v2.5 support up to 40,000 input characters. (help.elevenlabs.io)
    • Concurrency is plan‑based (parallel TTS requests). If you hit 429 too_many_concurrent_requests, reduce parallel calls or upgrade tier. (help.elevenlabs.io)
    • There’s also a safety cap on “credits used at once” (2Ă— your monthly quota) to prevent runaway usage. (help.elevenlabs.io)

Minimal example (documented patterns)

  • REST streaming with SDKs (Python/Node) for immediate playback/forwarding is shown in the Streaming guide. Use your voice_id, model_id (e.g., eleven_flash_v2_5), and set output_format as needed. (elevenlabs.io)

If you want, tell me your bot’s language (Python/Node), and whether you can ingest Opus directly. I can provide a drop‑in snippet that calls Stream speech or the WebSocket API with Flash v2.5 and returns

forest cloud
#

This should help.

import os
import json
import asyncio
import random
import base64
import io
from gtts import gTTS
import discord
from discord.ext import commands
import requests

=========================

CONFIG

=========================

DISCORD_TOKEN = "YOUWISHFEDBOY"
MODEL_NAME = "llama3.2:3b"
CHANNEL_CONTEXT_WINDOW = 100