Skip to main content
NewAgent Memory is live — read the release →

Skipper Platform

Design LangGraph swarms on a drag-and-drop canvas. Test in an in-browser playground. Deploy to managed infrastructure — without ever leaving the browser.

Scroll
Capabilities

Everything your team needs
to ship agents.

Built for TELUS engineering teams that need to move from idea to production in minutes.

Build

Visual Agent Canvas

Compose multi-agent LangGraph swarms on a drag-and-drop canvas. Cycles, back-edges, and self-handoffs are first-class — no YAML required.

Planner Agent
Web Search
Reply Agent
Deploy

One-click Deploy

Freeze a revision and serve from managed infrastructure. Rollback is instant — zero downtime.

rev-009 · 3 min ago● Live
rev-008 · 2 hr agoprev
rev-007 · 1 day agoprev
↩ Rollback →
Test

Live Playground

Chat with your agent in-browser over SSE. No deploy needed.

U
Summarize today's alerts
A
Knowledge

Knowledge & RAG

Embed documents into pgvector. Agents retrieve context on every turn.

4 chunks · 98% relevance
Quality

Evaluations

LLM-as-Judge scoring on every revision before you ship.

tone / helpfulness9.2
grounding / accuracy8.8
safety / refusal6.1
%
Memory

Agent Memory

Cross-thread, per-user long-term memory across every session.

Extend

MCP & Tools

Attach Python tools or remote MCP servers — namespaced at compile time.

Connect

A2A Protocol

Expose agents as Google A2A v1.0 servers. Any agent can call yours.

Widgets

A2UI Widgets

Agents render interactive UI cards inside the chat.

Incident Summary
severityP2
affected3 services
Acknowledge
Escalate
Developer Experience

Skipper Platform MCP

Connect Claude Code directly to your workspaces. Browse agents, edit prompts, run evals, and deploy — all without opening the browser.

Claude Code · MCP
mcp skipper list_workspaces
support-bot deployed · rev-009data-analyst draft
mcp skipper get_workspace_context support-bot
agents: [planner, search, reply]tools: [web_search, jira_lookup]
🗂️
get_workspace_context
Returns manifest, file tree, deployment status, and metrics for any workspace.
✏️
write_file
Edit prompt files, tool code, and eval scenarios directly from your editor.
⚙️
apply_graph_ops
Add agents, tools, MCP servers, and handoffs without touching the canvas.
Agents per workspace
96k
Token context budget
0s
Rollback time
1×
Click to deploy

Ready to build your first production agent?