Skip to content
AvelinLabs

AvelinLabs Examples

Complete executable examples are maintained in the official AvelinLabs API examples repository.

The repository is the executable companion to these docs. It contains complete cURL and Python workflows, payloads, sample responses, Postman assets, Customer Grounding files, MCP configuration, and a local interactive demo. The documentation explains the decision concepts; the repository demonstrates how to call them.

If you are using examples to test fit in a product workflow, follow Evaluate AvelinLabs in your workflow alongside the sample payloads and responses.

For practical strong, weak, ambiguous, vague, and noisy input comparisons, see Input quality examples.

During beta, onboarding is API-first: register through the Platform API, verify your email, create a runtime API key, and use that key on product endpoints.

Example Categories

Section titled “Example Categories”

Current public examples cover:

  • platform onboarding
  • runtime API key usage
  • job analysis
  • job classification
  • occupation intelligence
  • occupation alignment
  • skill evidence
  • confidence, uncertainty, and review-routing signals
  • weak-signal and ambiguity handling
  • labor market intelligence
  • market context where available
  • runtime authentication patterns
  • Customer Grounding examples for authorized Runtime API Key contracts
  • customer-operated MCP stdio configuration and tool calls
  • a local interactive Customer Grounding demo
  • API error handling
  • readiness checks

Current Public Example Files

Section titled “Current Public Example Files”

Customer Grounding Examples

Section titled “Customer Grounding Examples”

Customer Grounding examples apply when the Runtime API Key contract includes /api/v1/grounding. They cover the source lifecycle, text and supported file ingestion, grounded Role Intelligence, evidence, traces, errors, cleanup, and the customer-operated MCP adapter.

The controlled-beta example collection includes:

The examples do not imply production GA, connectors, OCR, vector retrieval, SDK support, a hosted MCP service, a production customer UI, or automated hiring decisions.

How to Read AvelinLabs Output

Section titled “How to Read AvelinLabs Output”

confidence

Section titled “confidence”
  • Meaning: Overall confidence signal for the top result.
  • Use it for: Routing and review, not as absolute truth.

trust_score

Section titled “trust_score”
  • Meaning: Trust-oriented score combining result strength and quality support.
  • Use it for: Deciding whether the selected output has enough evidence behind it.

uncertainty.total

Section titled “uncertainty.total”
  • Meaning: Combined uncertainty signal.
  • Use it for: Identifying inputs or matches that may need review.

decision.decision

Section titled “decision.decision”
  • Meaning: Routing label such as AUTO_ACCEPT, REVIEW, REJECT, or AMBIGUOUS.
  • Use it for: Deciding whether to display, review, reject, or request more input.

job_signals

Section titled “job_signals”
  • Meaning: Skill, tool, task, or domain signals detected from the input.
  • Use it for: Seeing what evidence the analysis used.

results[].onet_code

Section titled “results[].onet_code”
  • Meaning: Standard O*NET occupation code for a ranked candidate.
  • Use it for: Storing, comparing, or enriching the occupation consistently.

matching_skills

Section titled “matching_skills”
  • Meaning: Skills or capabilities supporting the selected match.
  • Use it for: Helping reviewers and client-facing teams explain why the result is relevant.

explanation.why_match

Section titled “explanation.why_match”
  • Meaning: Plain-English explanation for the candidate match.
  • Use it for: Reviewing and presenting the output in context.

For a fuller field-by-field walkthrough, read docs/annotated-job-analyze-response.md.

For practical field-by-field usage guidance in the public docs, see the Response Field Reference.

For practical batch and workflow measurements, see Metrics you can derive from AvelinLabs responses.

Controlled-Beta Guidance

Section titled “Controlled-Beta Guidance”

Use synthetic or explicitly approved customer-safe content. Store Runtime API Keys outside committed files and screenshots. The local demo is an evaluator tool, not a hosted product UI.

Contract Rule

Section titled “Contract Rule”

Examples should not invent fields, endpoints, or response shapes. If an API area is early-stage or subject to change, the example should say so directly.

Section titled “Related Guides”