together-pack

Gate Together AI client, batch, fine-tuning, and deployment changes with offline contract tests plus a protected bounded live lane. Use when adding CI for a Together-backed repository. Trigger with "Together CI", "test Together integration", or "Together contract tests".

Analyze and diagnose Together AI authentication, billing, request, model, throttling, overload, batch, fine-tuning, and endpoint failures from redacted evidence. Use when a Together integration fails or behaves inconsistently. Trigger with "Together error", "Together 429", or "Together model not found".

Prepare, submit, monitor, and disposition a Together AI fine-tuning job using SDK v2, validated training data, explicit cost approval, and separate deployment verification. Use when adapting a model to custom examples or preferences. Trigger with "Together fine-tune", "train a Together model", or "Together DPO job".

Run Together AI asynchronous batch inference from validated JSONL through upload, job polling, output/error download, and custom-id reconciliation. Use when bulk work can trade latency for lower cost. Trigger with "Together batch inference", "bulk Together requests", or "Together Batch API".

Reduce Together AI spend using measured token usage, live per-model prices, cached-input evidence, batch discounts, model evaluation, and dedicated break-even analysis. Use when forecasting or optimizing Together workloads. Trigger with "Together cost", "optimize Together spend", or "Together batch savings".

Deploy and roll back Together AI integrations across serverless inference or v2 Dedicated Model Inference with secret injection, health probes, traffic control, and cost shutdown. Use when releasing Together-backed services or dedicated models. Trigger with "deploy Together", "Together dedicated endpoint", or "Together rollout".

Build a bounded Together AI chat-completion probe with a current model selected from the live catalog, optional streaming, usage capture, and redacted evidence. Use when proving a new inference connection. Trigger with "Together hello world", "first Together request", or "stream Together chat".

Install Together AI SDK v2 and configure a project-scoped API key with least-privilege storage and a read-only model-list verification. Use when connecting a service or workstation to Together AI. Trigger with "Together auth", "install Together SDK", or "TOGETHER_API_KEY setup".

Develop Together AI integrations with a fake transport, recorded response shapes, deterministic assertions, and an opt-in bounded live probe. Use when iterating without spending tokens on every test. Trigger with "Together local dev", "mock Together API", or "test Together client".

Review a Together AI production release across model policy, auth, data controls, dynamic limits, retries, observability, cost, deprecations, asynchronous recovery, and rollback. Use when reviewing go-live or a material model change. Trigger with "Together production checklist", "Together go live", or "Together readiness review".

Analyze and control Together AI serverless concurrency from dynamic per-model request/token headers, bounded queues, jittered retries, and batch or dedicated alternatives. Use when preventing throttling or sizing inference traffic. Trigger with "Together rate limit", "Together 429", or "Together concurrency control".

Design a production Together AI service with a typed provider boundary, policy-based model routing, serverless and dedicated lanes, batch workers, telemetry, budgets, and reversible degradation. Use when defining the integration topology. Trigger with "Together architecture", "Together model gateway", or "design Together service".

Encapsulate Together AI SDK v2 behind a typed adapter with catalog-resolved models, bounded retries, streaming normalization, usage capture, and OpenAI-compatible migration seams. Use when building a reusable Together client layer. Trigger with "Together SDK pattern", "Together client wrapper", or "OpenAI compatibility on Together".

Secure Together AI integrations with project-scoped keys, environment isolation, prompt/output data controls, bounded model behavior, safe logging, rotation, and incident response. Use when threat-modeling or hardening Together usage. Trigger with "Together security", "protect Together API key", or "Together data controls".

Migrate Together AI Python SDK v1, OpenAI-compatible clients, deprecated models, or legacy dedicated endpoints with inventory, contract tests, canaries, and rollback. Use when upgrading Together dependencies or provider resources. Trigger with "upgrade Together SDK", "Together model migration", or "migrate Together endpoint v1".

Convert Together AI batch, fine-tuning, upload, and dedicated-deployment job states into idempotent internal events using bounded polling and an optional owned callback. Use when integrating asynchronous Together work. Trigger with "Together job events", "Together callback", or "poll Together status".