Built on
The platforms we build on
We are model- and vendor-agnostic. We pick the right enterprise runtime, orchestration layer, and model per client. and ship governed agents on top.
Vendor-agnosticEnterprise-gradePortable via MCP
Stack
Nine platforms, one operating layer
Enterprise runtime
Google Cloud
Google Vertex AI Agent Builder
Managed agent runtime on Google Cloud. Gemini models, Agent Development Kit (ADK), grounded retrieval, tool use, and enterprise governance.
Enterprise runtime
Google Cloud
Google Agentspace
Workspace-grade surface for enterprise agents. unified search, NotebookLM-style reasoning, and a managed agent gallery for employees.
Enterprise runtime
Microsoft
Microsoft Copilot Studio & Azure AI Foundry
Low-code agent authoring (Copilot Studio) on top of Azure AI Foundry's model catalogue, evals, and Responsible AI tooling.
Enterprise runtime
Amazon Web Services
AWS Bedrock Agents
Managed agents on Bedrock. multi-model (Claude, Llama, Titan, Mistral), Knowledge Bases for RAG, action groups for tool use.
Orchestration
LangChain
LangChain · LangGraph · LangSmith
Open orchestration stack. LangChain for composition, LangGraph for stateful multi-step graphs, LangSmith for tracing and evals.
Models
OpenAI
OpenAI Platform & Agents SDK
Frontier reasoning models (GPT-5 family), Responses API, Agents SDK, and built-in tools (web, file search, code interpreter).
Models
Anthropic
Anthropic Claude & MCP
Claude models for safety-first reasoning, plus Model Context Protocol (MCP). the open standard for connecting agents to tools and data.
Multi-agent
CrewAI
CrewAI
Role-based multi-agent framework. define crews of specialist agents with goals, tools, and processes, then orchestrate them.
Workflow glue
n8n / Zapier
n8n & Zapier Agents
Hundreds of pre-built connectors with an agent layer on top. fastest path to wire agents into the long tail of SaaS tools.
Not sure which stack fits?
We pick the runtime, model, and orchestration per client. based on data gravity, governance, and total cost.