v3.3.4

Fix syntax error

Full Changelog: v3.3.3...v3.3.4

v3.3.3

Full Changelog: v3.3.2...v3.3.3

v3.3.2

What's Changed

• Update Dockerfile to reduce image size by @DingtangSun0813 in #34
• Docker tagging improvements and python unused variable delete

New Contributors

• @DingtangSun0813 made their first contribution in #34

Full Changelog: v3.3.1...v3.3.2

v3.3.1

What's Changed

• v3.3 by @dbccccccc in #29
• Update Docker and release workflows by @dbccccccc in #30
• Improve audio combining and text splitting logic by @dbccccccc in #31

Full Changelog: v3.2.9...v3.3.1

v3.3.0-alpha5

Full Changelog: v3.3.0-alpha4...v3.3.0-alpha5

v3.3.0-alpha4

Full Changelog: v3.3.0-alpha3...v3.3.0-alpha4

v3.3.0-alpha3

v3.3.0-alpha3

Full Changelog: v3.3.0-alpha2...v3.3.0-alpha3

v3.3.0-alpha2

What's Changed

• Update Docker and release workflows by @dbccccccc in #30
• Improve audio combining and text splitting logic by @dbccccccc in #31

Full Changelog: v3.3.0-alpha...v3.3.0-alpha2

v3.3.0-alpha

Added

• Flask web app now spins up per-request TTS clients, streams responses, hashes API keys, and enforces simple rate limits.
• WebSocket handler tracks active tasks, supports cancellation, and streams base64 audio chunks for responsive UIs.
• Frontend metadata fetches use a cached API client, and CI now runs lint/test gates before packaging artifacts.

Changed

• Deterministic HTTP header generation no longer depends on network user-agent lookups.
• Default narration prompts are opt-in for both sync and async clients, and retry logic preserves original exceptions.
• Audio combining requires "pydub" for non-WAV formats and fails fast when the dependency is missing.

Fixed

• Async client retry payload handling copies request bodies per attempt to avoid mutation side effects.

Release v3.2.9: Refactor Docker build and enhance web playground

TTSFM v3.2.9

New release of TTSFM - Free Text-to-Speech API with OpenAI compatibility.

Installation

pip install ttsfm==3.2.9

Quick Start

from ttsfm import TTSClient

client = TTSClient()
response = client.generate_speech("Hello from TTSFM!")
response.save_to_file("hello")

Docker

docker run -p 8000:8000 dbcccc/ttsfm:latest

Features

• Completely free (uses openai.fm service)
• OpenAI-compatible API
• 11 voices available
• 6 audio formats (MP3, WAV, OPUS, AAC, FLAC, PCM)
• Async and sync clients
• Web interface included
• CLI tool available

Documentation

See README for full documentation.