Localized reading guide
duix-avatar-conversationPurpose, usage and source
Create a realtime conversational AI digital human via duix-cli: upload portrait (format + face check), select TTS voice (with preview), select language, optionally set name/greetings/profile, confirm custom-quota deduction, then generate and return conversation info + URL. Use when the user asks for realtime conversational avatar, custom digital human, talking avatar chat, interactive digital human, 实时对话数字人, 定制数字人对话, or mentions duix-avatar-conversation / duix-get-avatar-create-result.
Source retrieved:
- Source repository
- duixcom/duix-skills
- SKILL.md
duix-avatar-conversation/SKILL.md- Repository stars
- 12
- Repository updated
- 2026-08-12
- Categories
- Other
- Last collected
License: License file needs review
Workflow excerpts are reading aids, not executable instructions from Codez. No third-party Skill was run to produce this page.
The purpose and workflow passages below are upstream material, organized and translated for reading. They are distinct from the separately labeled Codez editorial analysis.
What this Skill does
Create a realtime conversational AI digital human via duix-cli: upload portrait (format + face check), select TTS voice (with preview), select language, optionally set name/greetings/profile, confirm custom-quota deduction, then generate and return conversation info + URL. Use when the user asks for realtime conversational avatar, custom digital human, talking avatar chat, interactive digital human, 实时对话数字人, 定制数字人对话, or mentions duix-avatar-conversation / duix-get-avatar-create-result.
How the upstream workflow works
How to interpret the response
When need_select is true: Extract data.options or data.skillPayload.data.options List every option. For each option, show index + label/value, and if it has preview audio (preview_url / url / exampleAudioUrl), include a clickable 试听 / Preview link in that same line Ask the user to listen (when a link exists) and then pick one
How to interpret the response
When need_confirm is true: Show CLI msg (preferred), e.g. 💡 定制次数确认 本次数字人对话生成需消耗 1 次定制次数，当前余额 X 次。确认提交请回复"是"，取消请回复"否"。 Wait for explicit user reply User says 是 / yes / y → continue to Step 6 with --confirm 是 (or --yes)
When to Use
Use this skill when the user wants a realtime conversational digital human / custom avatar / talking avatar chat.
Upstream passage · L23 ↗Requirements and setup
Prerequisites
Before every avatar operation, check the installed CLI against the official npm registry and upgrade it when it is not current. Do not continue until the upgrade succeeds and the installed version matches the registry version. Required environment variables (avatar commands):
Upstream passage · L27 ↗npm i duix-cli -g --registry=https://registry.npmjs.org/Limits and quality checks
How to read status (important)
Note: skill_code=100 is reused for TTS select / quota confirm / processing. Always check need_select, need_confirm, and task_status / task_id before deciding what 100 means.
Pitfalls
Follow this order only: image (+ check) → user TTS select (with preview) → language → optional persona → quota confirm → submit → generate Ask optional fields proactively, and allow "skip"
Upstream passage · L392 ↗Start with a safe trial
- Check the supported host and the repository installation instructions. Keep the Skill directory and its referenced scripts, assets and documentation together; SKILL.md alone may omit dependencies.
- Give the agent your real input, desired output and acceptance criteria. Invocation syntax and available tools depend on the host; this site does not assume a universal install command.
- Try the source workflow in a disposable project or on a copy. Review operations that change files, remote services or billable resources before allowing them.
- Validate the output against the source quality checks and your own requirements. A successful trigger or a high repository Star count does not establish a correct result.
Read further in the source
- Duix Skills - duix-avatar-conversation
- When to Use
- Prerequisites
- Agent Hard Rules (MUST)
- Guided Conversation Flow (Mandatory Order)
- Step 1: Upload image (required) + validation
- Step 2: Select TTS voice (required HARD GATE, with preview)
- How to interpret the response
- Step 3: Select language (default English, can skip)
- Step 4: Optional persona settings (ask proactively; each can skip)
- Step 5: Confirm custom-quota deduction before submit (required HARD GATE)
- How to interpret the response
- Step 6: Final summary and submit with
--confirm - Step 7: Poll result and return conversation info / failure
- How to read status (important)
- Anti-stuck rules for agents
- Success deliverable (required)
- Failure deliverable (required)
- Command Cheat Sheet
- Error Presentation
- Pitfalls
- For Automation Authors (
duix_run.sh/ scripts)
Related Skills
Translation and scope
This reading guide selects passages from the pinned upstream SKILL.md. Translation is assisted. It is not a complete manual or a claim of tested compatibility. Read the original and its referenced files before use.
This introduction refers to the pinned version below; current repository activity and collection time are shown separately.
- Pinned source
- Read the original ↗
- Source retrieved
- Source commit
e9cd894bf4ec1e91178d0af795d6ef3711e1a38f- Source SHA-256
f35e2907e291c70624dc17aaa2bfc415925146a42c569bc95c509d615a6191a5