Claude Skill MartBrowse skillsWhat is a Skill?
 
← Back to list

StepFun TTS (stepaudio-2.5-tts)

Generate emotionally expressive Chinese and Japanese speech with StepFun's stepaudio-2.5-tts using natural-language instructions and inline prosody directives.

Image & VideoIntermediate★ 1,323⑂ 212AI score 8/10Last updated: Aug 8, 2026

What it does

This skill wires Claude Code into StepFun's stepaudio-2.5-tts contextual TTS model. Instead of fixed emotion tags (voice_label, as in step-tts-2), you describe the delivery in natural language via the instruction field (≤200 chars) and drop inline () directives into the text for pauses, breath, emphasis, and mid-sentence mood pivots.

  • Single-line (1–500 chars) and long-form (up to a 1000-char hard cap) synthesis
  • JSONL batch generation for game/app voice lines
  • A/B comparison script across two model outputs
  • A full step-tts-2 → stepaudio-2.5-tts migration playbook
  • Documented handling for censorship_block errors plus mixed-model fallback

Who it's for

  • Game and app developers needing Chinese/Japanese voice lines
  • Teams migrating legacy step-tts-2 pipelines to the 2.5 model
  • Audio content creators who need whispers, sobs, playfulness, and other nuance
  • Engineers calling the StepFun API directly who want fast error diagnosis

Examples

  1. One emotive line: python3 scripts/tts_generate.py --text "你好" --out /tmp/hello.mp3 --instruction "温暖的希望感" for a warm, hopeful greeting.
  2. Mid-sentence emotion shift: write "(试探着问)你好吗？(开心地)太好了！" — the parenthesised text acts as stage direction and is never spoken aloud.
  3. Batch of 29 game lines: run --batch lines.jsonl; if 2 lines return censorship_block, log the IDs, keep the batch running, and regenerate just those two through step-tts-2.

· · · Install guide · · ·

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the daymade-audio/stepfun-tts folder from the GitHub repo daymade/claude-code-skills into my ~/.claude/skills/stepfun-tts/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/daymade/claude-code-skills.git && mkdir -p ~/.claude/skills && cp -r claude-code-skills/daymade-audio/stepfun-tts ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal and clone the repo: git clone https://github.com/daymade/claude-code-skills.git
  2. Create the skills directory: mkdir -p ~/.claude/skills
  3. Copy this skill in: cp -r claude-code-skills/daymade-audio/stepfun-tts ~/.claude/skills/
  4. Sign up at https://platform.stepfun.com/ and create a key under API Keys. Choose a Normal key — Plan keys are text-only and fail silently on audio endpoints.
  5. Export the key: export STEPFUN_API_KEY="your-key-here" (add it to ~/.zshrc or ~/.bashrc to persist).
  6. Restart Claude Code and ask something like "synthesize this Chinese line with StepFun TTS, whispered".
  7. Smoke test: run python3 ~/.claude/skills/stepfun-tts/scripts/tts_generate.py --text "你好" --out /tmp/hello.mp3 and play the resulting mp3.
View source on GitHub ↗License: MIT