Crescendo Lab
繁體中文 Book a demo

Crescendo Agent Lab · Enterprise AI agent platform

Build your company's AI assistants, connect your systems, keep IT in control.

Crescendo Agent Lab is an enterprise AI agent platform: build AI assistants in one workspace, connect them to GitHub, Notion, Jira, your ERP and more, and let employees use them right inside Slack, Teams and LINE. Assistants act only with each employee's own permissions, and every model call is capped, metered and exportable for audit.

  1. 1Build assistantsInstructions, knowledge, skills, versions and evals
  2. 2Connect your systems24 one-click integrations, any MCP server
  3. 3Use them in Slack, Teams and LINENo new tool for employees
Book a demo See how it works

In 45 seconds

One request, from ask to audit trail

Watch one agent tool call pass identity, approval and budget, then land as a single row in the ledger.

Integrations

Connect the tools your company already runs on

24 integrations are one click away in the console, any remote MCP server can be added by its URL, and Zapier bridges thousands more apps. Every connection runs through the same identity, approval and audit trail.

  • 24integrations you add in one click
  • Anyremote MCP server, added by its URL
  • Thousandsof 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 authorization or one 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

How it works

Governance built into the platform, not bolted on by consultants

These three mechanisms are in the platform's code. Every workspace has them from day one.

effective = user ∩ assistant ∩ tool User's own permission what they already see in ERP Assistant's capability which tools it was granted Tool's own permission what the credential allows Effective access

Delegated identity:
assistants act as the person asking

For every tool call, the platform computes an intersection: what this person can already do in the downstream system, what this assistant was granted, and what the tool itself allows. It runs only if all three agree. Your existing permissions are the rules.

  • Knowledge retrieval also runs as the asker, so documents they cannot see are never cited
  • A connector can be a shared organization connection or the member's own authorization
  • Permission is checked again right before execution, so removed members stop acting
Model proposes a tool call Three checks tool · rules · screening Waits for a person awaitingApproval Slack · #procurement Wants to call erp.inventory.lookup Runs with Sam Lin's permissions Approve Deny Runs and is recorded who asked · who approved · as whom The same card works in the console Read-only tools an owner marks as automatic can skip the wait

Approval where work happens:
tool calls pause for a person

By default every tool call the model proposes waits for someone to approve it — in the console or on a Slack card, one state machine behind both. After approval the platform checks permission again before running, and a call that already ran is never run twice.

  • Business rules can require approval for certain tools, or forbid certain actions outright
  • When content screening escalates a turn, every tool call in it needs approval
  • Read-only tools an owner marks automatic can run unattended, and one switch turns that off
Workspace monthly budget 62% 70% 90% Under 70% the model the member chose Claude Opus by default 70% to 90% steps down to a cheaper model and says so on the reply Past the ceiling refused with a reason never borrows next month Ledger · one row per model call 10:42 procurement · sam.lin · claude-opus · 3,120 tok · chat 10:43 procurement · sam.lin · claude-opus · 1,840 tok · tool 11:00 weekly-report · auto · claude-haiku · 22,400 tok · schedule 11:05 kb-index · system · claude-haiku · 8,900 tok · index

Budgets that hold:
ceilings per workspace and per team

Every model call takes one path: check the budget, call the model, write the ledger. Near the ceiling, replies move to a cheaper model and tell the reader. Past it, requests are refused with a reason. A reply where the member named a model is never silently swapped — it is refused instead.

  • Monthly workspace budgets, rolling-window limits, and per-team ceilings
  • Each call records who, which assistant, which model, how many tokens, and why
  • The usage report answers "why was this reply so expensive"

Security and deployment

What IT checks before go-live is already built in

Where data lives, who can sign in, and how to trace an incident are part of the platform, not settings added during rollout.

Data and isolation

  • Runs on Google Cloud in Taiwan (asia-east1), with a tenant boundary per workspace
  • A workspace can have its own dedicated Cloud SQL database
  • Connector credentials sealed with AES-256-GCM, with short-lived grants per use
  • On-premises systems connect through a relay, so data does not have to move to the cloud first

Identity and audit

  • Enterprise SSO with SAML and OIDC, with directory groups mapped to workspace roles
  • Delegated identity, tool approvals, and token budgets with step-down in every workspace
  • Tool approvals, membership changes, and credential use are recorded, and owners can export them
  • Slack, Microsoft Teams, and LINE share one identity, budget, and audit trail

Our first customer is us

Franky runs on this platform

Franky is the AI coworker Crescendo Lab's staff @mention in Slack, and every answer cites a source. The identity, permissions, budgets, approvals, and audit behind it are Agent Lab — now a multi-tenant platform any company can open a workspace on.

Crescendo Lab · Franky internal adoption, last 30 days
Meet Franky

Book a demo — bring your IT team

  • 45 minutes walking an approval and a budget through your own scenario
  • Straight answers on data residency, permissions, and cost control
  • If it fits, a workspace for your team the same week
Book a demo

Optional — helps us schedule faster