Zapier-grade workflows + Manus-style autonomous agents

Describe the outcome. Operon runs the work.

Connect Google Calendar, Drive, Gmail, Sheets, Docs, Slack, and GitHub. Operon generates a validated execution plan, verifies permissions, and pauses for your approval before sensitive actions.

Free plan includes 500 runs a month · Powered by VNAI Pro · Encrypted Vault OAuth

Interactive PreviewNeeds approval
Natural-Language Outcome PromptScheduled · 0 8 * * 1-5

“Every weekday at 8 AM, summarize my upcoming Google Calendar events and email me a briefing.”

Verified Scopes:Google CalendarGmail
  1. 0.3s

    calendar_list_events({ calendarId: 'primary', maxResults: 20 })

    Plan note

    Query primary Google Calendar for events occurring in the next 24 hours.

    Tool input

    calendarId: primary
    window: Next 24h (America/New_York)

    Tool output

    {"count": 6, "events": ["Q4 Architecture Sync", "Customer Renewal Call", "Design Review"]}
  2. 0.5s

    vnaipro.chat.completions({ model: 'claude-sonnet-5-5' })

  3. Needs approval

    gmail_send_message({ to: 'founder@company.com' }) · approval gate

    Waiting for human approval

    Review the tool call parameters before allowing this step to execute.

Click any step to inspect validated inputs and JSON outputs.

How it works

Three steps from prompt to verified result.

Every run follows the same transparent trace so you always know what ran, what is waiting on you, and why.

  1. Step 01

    Describe

    Write what you want done in plain English or pick a trigger on the workflow canvas. Operon identifies which connected apps and tools are needed.

    Prompt or webhook trigger → tool selection
  2. Step 02

    Review the plan

    Before any step executes, Operon lays out the ordered trace. You can inspect arguments, adjust conditions, or mark sensitive actions with an approval gate.

    Full step trace visible before execution
  3. Step 03

    Approve and run

    Read-only steps run automatically. When a run reaches an approval gate—like sending an email or issuing a refund—it pauses until you approve or reject.

    One-click approve or reject from desktop or phone

Two ways to automate

Ask an agent or build a workflow.

Use an agent when the task requires reasoning across unstructured data. Use a workflow when you want the exact same sequence to run on a schedule or webhook.

Goal-driven

Ask an agent

Type a goal in plain English. The agent plans the necessary tool calls, adapts to the data it finds, and pauses when it reaches an action that changes external state.

“Summarize today's urgent support emails and draft replies.”
1. Read Gmail→2. Draft replies→3. Approve send
  • Best for research, triage, and one-off tasks
  • Streams reasoning and tool outputs live
Deterministic DAG

Build a workflow

Connect triggers, API calls, AI steps, and approval nodes on a visual canvas—or view and edit the same flow as an ordered list on mobile.

Trigger: Stripe webhook (charge.dispute.created)
Postgres query──AI summary──Slack alert
  • Best for recurring schedules and webhook pipelines
  • Canvas editor on desktop, list inspector on mobile

Security & control

Encrypted credentials and human sign-off before sensitive actions.

Operon is designed so the language model never sees your raw API keys and never executes high-impact writes without your permission.

Encrypted credential vault

OAuth tokens and API keys are encrypted at rest with AES-256. Credentials are injected only by the server runtime when a step executes—never included in model prompts or logs.

Human approval gates

Mark any step as requiring approval. When a run reaches that step, execution pauses and shows you the exact payload so you can approve or reject it.

Complete step-by-step audit trail

Every run records its inputs, tool calls, approvals, durations, and outputs. Row-level security isolates every workspace at the database layer.

We never train models on your workflows, run history, or credentials.
Read our data manifest & privacy policy →

Pricing

Start free with 500 runs a month.

Every account starts on the Free plan with 500 monthly runs. Runs pause cleanly at your limit so you never receive a surprise overage bill.

Free

Default plan

Everything you need to run agents, build workflows, and verify human approval gates.

$0per month

500 runs / month

  • 500 agent and workflow runs per month
  • Visual workflow canvas and live trace timeline
  • Human approval gates before sensitive actions
  • Encrypted credential vault (AES-256)
  • All built-in app integrations
  • 30-day run history and step logs

Pro

Editable placeholder

Higher monthly run allowances and extended log retention for active builders.

Custom / TBDconfigurable tier

Configurable run allowance

  • Configurable monthly run allowance
  • Unlimited active workflows and schedules
  • Isolated Python and browser tool steps
  • 90-day run history and audit logs
  • Webhook triggers and REST API access

Team

Editable placeholder

Shared workspaces, role-based approval policies, and centralized credential vaults.

Custom / TBDconfigurable tier

Pooled team run allowance

  • Shared workspace and pooled monthly runs
  • Multi-member approval gates
  • Shared encrypted vault permissions
  • Audit log export and SSO configuration
  • Dedicated support channel

FAQ

Common questions, answered plainly.

Need more detail on security or self-hosting? Email us at support@operon.app.

Read-only steps (such as reading messages or querying a database) can run automatically. Any step marked with an approval gate—such as sending an email, issuing a refund, or updating a record—pauses the run and waits for you to approve or reject the exact payload.

Run your first agent trace in under a minute.

Start on the Free plan with 500 runs per month. Connect your apps when you are ready.