# Move AI agents from demos into real enterprise work.

> Crescendo Agent Lab is one enterprise platform for AI agents: entry points, decisions, system access, deployment and governance, from demo to real work.

Web page: https://agentlab.cresclab.com/en  
繁體中文 (Markdown): https://agentlab.cresclab.com/index.md

Crescendo Agent Lab · Enterprise AI agent platform

Agent Lab brings entry, decisions, connections, deployment, and governance into one enterprise platform.

[Sign up free](https://app.agentlab.cresclab.com/?lang=en) [Book a demo](https://agentlab.cresclab.com/en#demo)

Want the tour first? [See how Agent Lab works](https://agentlab.cresclab.com/en#modules).

Agent Lab platform architecture: a digital employee works across Channel, Decision, Connect, and Deployment, governed by Identity, Policy, Approval, Usage, and Audit.

The basics

## Crescendo Agent Lab: an enterprise AI agent platform from Taiwan

Crescendo Agent Lab is the enterprise AI agent platform from Crescendo Lab in Taiwan: people hand work to agents in the web console and Slack, and IT manages permissions, budgets and records in one place.

**Choosing one**

[How to choose an enterprise AI agent platform: six questions for IT](https://agentlab.cresclab.com/en/platform#choose)

**Modules**

Channel / SDK, AI Gateway, Access & Connect, Deployment

**Permissions**

Agents act only with the asker's own permissions

**Cost**

Workspace and team budgets are hard caps; calls past them are refused

**Records**

Every model call, approval and credential use logged and exportable

**Where it runs**

Google Cloud in Taiwan; hybrid or on-premises per project

Four core products

## An agent needs four capabilities to work inside an enterprise

Four products cover the enterprise agent lifecycle.

### [Channel / SDK](https://agentlab.cresclab.com/en/channels)

Put agents in Web, App, LINE, Slack, Teams, or your own API.

Let agents be used →

### [AI Gateway](https://agentlab.cresclab.com/en/gateway)

Unify model routing, budgets, usage, and cost.

Let agents decide →

### [Access & Connect](https://agentlab.cresclab.com/en#access-connect)

Connect APIs, MCP, databases, and SaaS with permissions and audit.

Let agents act safely →

### [Deployment](https://agentlab.cresclab.com/en/deploy)

Run and monitor agents in cloud, private cloud, or on-prem.

Let agents run →

In 45 seconds

## Four modules make agents work. One platform keeps them under control.

Front door, brain, hands and runtime, held up by one layer of governance.

Channel and decisioning

## Carry one identity, policy, and usage trail through every interaction

Channel / SDK brings agents to users; AI Gateway governs models, tools, and cost.

01

### Channel / SDK

Put existing agents where employees and customers already work.

- Web
- App
- LINE
- Slack
- Teams
- API / SDK

[Explore Channel / SDK →](https://agentlab.cresclab.com/en/channels)

02

### AI Gateway

Every model call passes one point: pick the model, reserve the budget, write the ledger.

- Auto routing
- OpenAI / Anthropic API
- Claude Code · Codex
- Budget
- Usage
- Cost

[See how the AI Gateway works →](https://agentlab.cresclab.com/en/gateway)

03 · Access & Connect

## Connect systems and make every agent's authority explicit

Add 24 integrations in one click; reach anything else through remote MCP or Zapier.

- **24** integrations, one click each
- **Any** remote MCP server, by URL
- **Thousands** of apps through Zapier

- GitHub
- Notion
- Linear
- Atlassian
- Asana
- Figma
- Sentry
- GitLab
- Intercom
- monday.com
- Airtable
- Zapier
- Stripe
- PayPal
- Shopify Storefront
- Webflow
- Wix
- Supabase
- Vercel
- Postman
- Hugging Face
- Cloudflare Docs
- DeepWiki
- Context7
- Any MCP serverPaste the URL, pick the auth

Every connection is governed the same way

- The member's own or a shared organization connection
- OAuth, bearer token or client credentials
- Credentials sealed with AES-256-GCM
- On-premises systems through a relay
- Slack, Teams and LINE share one identity and audit trail

04 · Deployment

## Deploy what your AI built, from the same chat you built it in.

Ship the site, app and PostgreSQL you built with Claude Code, Codex or another MCP client from one sentence in the chat, with a preview for every version.

[See how Deploy works](https://agentlab.cresclab.com/en/deploy) [Moving from another host →](https://agentlab.cresclab.com/en/deploy/migrate)

Agent Lab Platform

## Governance is not an add-on. It is the shared foundation for every agent.

Every use, decision, connection, and execution gets an identity, a rule, and a record.

- Identity, roles, and delegated access
- Policy, business rules, and human approval
- Usage, budgets, and cost control
- Audit, monitoring, cloud, hybrid, and on-premises

[Explore the Agent Lab Platform](https://agentlab.cresclab.com/en/platform)

Use Cases

## Turn platform capabilities into digital employees that get work done

Franky and DAAC are both built on Agent Lab.

Digital Employee

### Franky

A digital employee in Slack that answers with sources and works across systems with each coworker's permissions.

[Visit Franky's dedicated page →](https://franky.cresclab.com/)

Data Agent

### DAAC

Turns MAAC, CAAC, and external data into insight, predictions, and next actions.

[Explore DAAC on the main site ↗](https://www.cresclab.com/en/product/daac)

Docs

## One documentation entry point for guides, APIs, MCP, and SDKs

Quickstarts, product guides, and technical references.

Docs

### [Agent Lab Docs](https://agentlab.cresclab.com/docs/en)

Guides, API and MCP references, and the Channel SDK.

Open Docs →

FAQ

## Crescendo Agent Lab questions

**What is Crescendo Agent Lab?**

The enterprise AI agent platform from Crescendo Lab, Taiwan. Four modules (Channel / SDK, AI Gateway, Access & Connect, Deployment) let agents do the work, and one set of identity, budgets, approvals and audit keeps them in line.

**Is it the same as ServiceNow AgentLab or other products named AgentLab?**

No. Crescendo Agent Lab is made by Crescendo Lab and is unrelated to ServiceNow AgentLab, Resolve AgentLab or the agentlab packages on PyPI and npm.

**What should IT have in place before rolling out AI agents?**

Three things: agents that cannot overreach, a token bill with a hard cap, and a record of who did what. Agent Lab acts only with the asker's own permissions, holds the budget before each call, and records every model call and approval. [How to choose an enterprise AI agent platform](https://agentlab.cresclab.com/en/platform#choose)

**Where do employees use the agents?**

In Slack, Microsoft Teams, LINE group task intake or the web console, and through MCP and the Agent API in your own systems. Every entry point shares one identity, budget and record.

**Where is data kept? Can it run on-premises?**

The cloud edition runs on Google Cloud in Taiwan (asia-east1) with one tenant per workspace. Systems inside your network connect through a relay that only dials out; the whole platform in your data center is scoped per project.

**Can we try it first?**

Yes. Sign up to open a workspace yourself, or book a demo to walk one of your scenarios through approvals and budgets.

## Sign up free, or book a demo

[Sign up free](https://app.agentlab.cresclab.com/?lang=en)

Or book a demo and bring one real agent scenario:

- 45 minutes on an approval and a budget in your scenario
- On-prem or hybrid? We'll look at your environment
- Straight answers on data, permissions, and cost
- If it fits, a workspace the same week