betalyra
Official@betalyra · Portugal
Software & AI made at the Atlantic
Agent Skills by betalyra
Showing 17 vetted skills indexed across 1 GitHub repositories.
betalyraeffect-uai-tool-call-approval
Gates sensitive tool calls behind human approval verdicts before execution.
betalyraeffect-uai-multi-model-fallback
Retry AI conversations with alternate provider tiers on retryable failures.
betalyraeffect-uai-model-council
Coordinate multiple AI models to answer prompts and select a winner via cross-model scoring.
betalyraeffect-uai-streaming-structured-output
Decode and stream schema-validated JSON objects from model output as JSONL.
betalyraeffect-uai-basic-usage
Stream model turns, invoke tools, and aggregate outputs into a final answer.
betalyraeffect-uai-pause-resume
Pause and resume an in-flight agent loop using Latch primitives.
betalyraeffect-uai-structured-output
Return typed JSON objects validated locally against an Effect Schema.
betalyraeffect-uai-multi-model-compare
Send one prompt to multiple providers and stream tagged outputs.
betalyraeffect-uai-model-retry
Retry transient model failures with exponential backoff and jitter.
betalyraeffect-uai-modify-output-stream
Format a Stream<TurnEvent> for transport via SSE or JSONL.
betalyraeffect-uai-migrate
Apply per-version rename tables and behavior-change rules to migrate effect-uai call sites.
betalyraeffect-uai-mid-stream-abort
Cancel in-flight model turns using Effect's Stream.interruptWhen and Deferred signaling.
betalyraeffect-uai-auto-compaction
Summarize older conversation turns to stay within token budgets.
betalyraeffect-uai
Builds AI agent loops with Effect-uai primitives like loop and streamTurn.
betalyraeffect-uai-embedding
Convert text and images into vector embeddings for semantic search and retrieval.
betalyraeffect-uai-agentic-loop
Buffer user messages and debounce typing into batch turns for chat agents.
betalyraeffect-uai-streaming-tool-output
Stream intermediate tool events and reduce them into a single structured final result.
Frequently Asked Questions About betalyra
FAQPage SchemaWhat specific tasks can be performed using these primitives?▼
These primitives enable structured JSON streaming, multi-model consensus scoring, human-in-the-loop execution gating, and resilient retry logic for transient model failures. You can also perform semantic vector embedding generation and manage complex stateful conversation loops with built-in token budget compaction.
Which engineering personas benefit from this library?▼
This library is designed for TypeScript engineers building production-grade distributed systems who require strict type safety and functional error handling. It is ideal for developers integrating complex model interactions into existing Effect-based architectures who need granular control over streaming, state, and execution flow.
What are the prerequisites for implementing these capabilities?▼
Implementation requires a project environment configured with the Effect ecosystem. Developers must define schemas for structured output validation and configure provider credentials for the models being orchestrated. The library integrates directly into existing Effect-based codebases to manage stream lifecycle and deferred signaling.