| name | voyager-help |
| description | Get help with Voyager SDK for Axelera AI hardware. Use when the user asks questions about Voyager SDK, Axelera hardware, pipelines, InferenceStream API, or model zoo. Explanation-only answers; no code changes or pipeline runs. |
| argument-hint | <question about Voyager SDK> |
| allowed-tools | Read, Bash, Glob, Grep, mcp__voyager__* |
Voyager SDK Help
Get help with Voyager SDK for Axelera AI hardware
Important Context
Answer with explanations only; do not change code or run pipelines.
Instructions
Answer questions about Voyager SDK: $ARGUMENTS
Step 1: Data Source & Environment Selection
{{INCLUDE common/voyager-sdk-setup.md}}
Step 2: Answer the Question
After gathering information from the selected source:
-
Synthesize Response
- Always frame answers in context of Axelera AI hardware
- Provide Voyager SDK-specific code examples only after checking SDK/RAG evidence
- Reference specific SDK components (InferenceStream, deploy.py, inference.py)
- Distinguish
documented, observed locally, and not validated here
- Include a concise source note naming the SDK file, command output, or RAG result used
-
Include Practical Examples
- Show actual command-line usage
- Include YAML pipeline snippets when relevant
- Reference model zoo locations (
ax_models/zoo/)
-
Hardware Context
- Mention detected Axelera device when relevant
- Include AIPU core configuration options
- Reference hardware capabilities such as vaapi, opencl, or opengl only when
.voyager-runtime.json or SDK documentation supports them for this host
Common Topics
| Topic | RAG Query | Key Files |
|---|
| Pipeline YAML | "pipeline YAML structure" | ax_models/zoo/**/*.yaml |
| InferenceStream | "InferenceStream API" | axelera/app/stream.py |
| Deployment | "model deployment compile" | deploy.py, axelera/app/compile.py |
| Input Sources | "USB RTSP video sources" | axelera/app/config.py |
| Custom Operators | "custom operator pipeline" | axelera/app/operators/ |
| Trackers | "ByteTrack OC-SORT" | trackers/ |
Example Usage
voyager-help: How do I set up multi-stream inference?
voyager-help: What is the difference between the gst and torch-aipu pipes?
voyager-help: How to create a custom post-processor?
voyager-help: Explain the pipeline YAML structure
voyager-help: How to optimize for Axelera PCIe with 4 cores?
