Tools for builders

Distill production agent traffic into smaller fine-tuned models.

Pipelines for retrieval-heavy agent workloads.

Hosted agent runtime with native function calling and code execution.

Type-safe agent framework from the team behind Pydantic.

Conversational multi-agent framework with strong reasoning patterns.

Role-based multi-agent framework with declarative crew definitions.

Durable workflows and step functions for agents.

Stateful, graph-based orchestration for LLM workflows with deterministic checkpoints.

Durable workflow engine increasingly used for long-running agents.

Long-term memory primitive: hierarchical context with explicit recall calls.

Reference servers and clients for the Model Context Protocol.

Hosted, isolated browsers for agent automation with session replay.

Cloud sandboxes for code-running AI agents.

Serverless infra for agent workloads — sandboxes, GPUs, schedules.

OpenTelemetry-native LLM observability and evaluation.

Lightweight LLM observability with a proxy-first model.

Open-source observability for LLM and agent applications.

Tracing, evals, and experiment tracking unified.

Open-source coding-agent IDE extension for VS Code and JetBrains.

No-code agent builder for business operations workflows.