Skip to content

Capabilities

Async Execution

Batch API and background mode for long-running or bulk inference

Read this when
  • Processing a large queue of independent prompts
  • Running deep-research / long agentic flows

Batch API

Async bulk inference. Submit a JSONL of requests, run within a window (typically 24h, often <1h), retrieve results at ~50% of synchronous cost.

ProviderAPIpi-goNotes
Anthropic✅ Message Batches; up to 300k output tokens with output-300k-2026-03-24 betanot wired
OpenAI Chat✅ Batch API; 50% discountnot wired
OpenAI Responsesnot wired
Google Gemini✅ via Vertex AI batch predictionsnot wired
Claude CLI

Background Mode

Long-running async generation: submit, poll status, retrieve.

ProviderAPIpi-goNotes
Anthropic❌ (use Batch instead)
OpenAI Chat
OpenAI Responsesbackground: true (preview)not wired
Google GeminiLive API is realtime, not background

Provider Documentation

pi-go Gaps

  • No BatchProvider capability interface; no JSONL builder, submit/poll/retrieve helpers.
  • No BackgroundProvider capability interface.
  • Both belong as optional capability interfaces alongside ImageProvider and ObjectProvider.
  • Without these, callers must use synchronous calls even for embarrassingly parallel workloads — forfeiting the ~50% Batch discount.