The AI workspace for businesses and the agencies that serve them.
Answer chats and calls with AI assistants, build real apps by describing them, capture and follow up leads, ask for reviews, and run research, content and reporting, for your own business or every client you manage. Anything that leaves the building waits for your approval.
7-day trial with a card on file and 300 usage credits. Then $49.99/month unless you cancel.
- Everything classic Versabot did, still here
- Approval before anything is sent
- GPT or Claude, or your own key
Follow-up email · waiting for approval
Exact contentWorks with: OpenAI GPT models, Claude Haiku 5.5, Claude Sonnet 5.5, Claude Opus 5.5, SMTP & IMAP mailboxes, LinkedIn, Website logins, REST APIs, MCP clients, Bring your own key
Platform
Everything your business runs on, around one agent
Assistants, phone, apps, lead capture, pipelines, outreach, dashboards and reporting share one workspace, one permission model and one evidence trail.
One agent, every client
Threads, projects, search and notifications around an agent that researches, drafts, browses and reports.
KPI dashboards
Live metrics with drilldowns that respect every role.
Forms that publish safely
Branching, consent capture and immutable versions.
Sandboxed browser agent
Real Chromium, read-only by default, with screenshots as evidence.
Learn more →Contacts & custom fields
Per-client pipelines, tags, typed fields and automatic de-duplication.
Learn more →Scheduled agents
Recurring checks and report drafts with a budget per run.
Learn more →Client portal
Publish reviewed reports with run evidence to each client.
Learn more →Honest review requests
Sent automatically after completed jobs, with a website review popup. Never gated.
Learn more →MCP server
21 scoped agency tools for external agents.
Learn more →How it works
From request to reviewed delivery
The agent does the legwork. You stay in charge of what goes out and what clients see.
- 1
Set up your business (or your clients)
Run Versabot for your own business, or add the clients you serve with scoped staff roles. Connect a mailbox, phone number, LinkedIn or Cloudflare in secure forms; credentials never go in chat and are stored encrypted.
- 2
Ask for the work in a thread
Threads keep context per client and project. Pick the model and reasoning effort for each thread: OpenAI GPT models or Claude Haiku, Sonnet and Opus.
- 3
Approve what leaves the building
Emails, review requests, interactive browser actions and code runs wait for explicit approval of the exact content. Editing a draft resets its approval.
- 4
Deliver and measure
Publish reports to a client portal with the run evidence behind them, and tie results to verified receipts rather than estimates.
Capabilities
What the agent can do today, labeled honestly
Each capability shows its real status, so you know what works now, what needs your own account, and what is still coming.
- Live
- released and verified in production
- New
- newly released in the workspace
- Needs your account
- works after you connect your own provider account
- Test mode
- runs end to end, but no real money moves
- Not yet available
- on the roadmap, not available today
Agent threads and projects
LiveApp Builder
NewAI assistants and chatbots
LiveAI phone and voice agents
LiveAI SDR and appointment booking
LiveResearch profiles
LiveContent studio
LiveSandboxed browser runs
LiveApproved code execution
LiveSchedules
LiveClient reports and portal
LiveContacts and custom fields
NewWebinars
NewForms and KPI dashboards
NewEmail outreach
Needs your accountOptional reply handling
Needs your accountReview requests
Needs your accountPerformance fees
Needs your accountSMS and WhatsApp
Needs your accountLinkedIn posting
Needs your accountAd platforms
Not yet availableBy the numbers
Measured, scoped and auditable by design
6
roles, from owner to client viewer
$0.01
per credit, measured from actual cost
0
platform LLM credits with your own key
21
agency tools over MCP
Accountability
Built so you can stand behind the work
- Exact-content approval for outgoing email, review requests, interactive browser steps and code runs.
- Six roles (owner, admin, member, auditor, client admin, client viewer), with staff restricted to the clients they serve.
- Browser runs keep steps, visited URLs and screenshots. Reports can cite that evidence.
- Credits are reserved before costly work and settled on measured usage. Unused holds are returned.
- Interrupted or uncertain actions are flagged for review and never silently replayed.
What we don't claim
- Versabot is not an all-in-one CRM, funnel builder and phone-system replacement. Its contacts and forms are deliberately lightweight.
- Attribution counts verified receipts only. It does not prove causal lift or every offline conversion.
- SMS needs your own A2P 10DLC-registered number, and WhatsApp needs an approved sender and template. We don't bypass carrier or Meta approval.
- Social posting covers LinkedIn today. Ad publishing to Google and Meta is not available yet.
- Google Business review monitoring and reply posting are not connected yet.
- Performance fees are invoiced from your own Stripe account. Versabot never holds client money.
Who it's for
For businesses, and the agencies that serve them
Local and service businesses
Marketing and growth agencies
Independent operators
Teams building for clients
Where it fits
One AI workspace, honest about its edges
Versabot goes deep on accountable AI work, assistants, phone and apps, and stays light on CRM. Many teams run it alongside the tools they already have.
| Versabot | General chat assistants | All-in-one agency suites | |
|---|---|---|---|
| Does real multi-step work with tools | ✓ | Partly | Partly |
| Approval of exact content before sending | ✓ | — | Partly |
| Per-client data, roles and portal | ✓ | — | ✓ |
| Sandboxed browser with run evidence | ✓ | — | — |
| Full CRM, funnels, phone and SMS | — | — | ✓ |
Detailed comparisons: vs GoHighLevel, vs ChatGPT and vs Zapier.
Models and keys
Your choice of model. Usage measured, not guessed.
Pick the model per thread
LiveBring your own key
LiveClassic Versabot, all of it
LivePricing
Start with a trial. Pay for measured usage.
One credit equals $0.01 of usage, measured from actual token and provider cost. New accounts start on Personal with a 7-day trial.
Personal
$49.99/month
500 usage credits per paid month
Pro
$99.99/month
1,500 usage credits per paid month
Business
$399.99/month
10,000 usage credits per paid month
Guides
Learn how businesses and agencies run work with AI
Fundamentals · 6 min
What is an AI agent workspace?
An AI agent workspace pairs a capable model with tools, memory, permissions and approvals so it can do real work. Here's how it differs from a chat window.
Fundamentals · 5 min
AI agent vs chatbot: what's the difference?
Chatbots answer; agents act. Learn the practical differences between AI agents and chatbots, when each fits, and what to check before you deploy one.
Operations · 6 min
Human-in-the-loop AI: how approvals work
How to design approvals for AI agents: what should need a person's sign-off, exact-content approval, editing resets, schedules and keeping an audit trail.
Reputation · 5 min
Review gating: why to avoid it
Review gating means asking only happy customers for reviews. Google prohibits it. Here's what counts as gating and how to ask for reviews the right way.
Pricing · 5 min
BYOK and usage-based AI pricing explained
How usage-based AI pricing works, what credits measure, and when bringing your own API key (BYOK) saves money. A plain-language guide for agencies.
Developers · 6 min
What is an MCP server? A practical guide
The Model Context Protocol (MCP) lets AI agents use tools from other systems. Learn how MCP servers work, how scoped tokens keep them safe, and common uses.
FAQ
Common questions
Does Versabot include a CRM?
A lightweight one. Each client gets contacts with pipeline stages, tags and your custom fields, fed by versioned lead forms and the agent, plus KPI dashboards. It is not a full CRM with deals, calling and SMS.
Is Versabot still a chatbot and voice-agent builder?
Yes. Everything classic Versabot did is still here: AI assistants and website chat, knowledge bases, plugins, voice and phone agents, the SDR and appointment booking. They now sit inside one agent workspace alongside research, content, the App Builder, outreach, reviews and reporting.
Will the agent send emails or publish anything on its own?
Not by default. Outgoing email and review requests require approval of the exact content, and reports stay drafts until someone publishes them. Automatic replies exist only if you turn them on for a specific connection, with named senders, a daily cap and an expiry.
How is usage priced?
Hosted AI usage is measured from actual token and provider cost and charged in credits, where one credit is $0.01. New accounts start with a 7-day, card-required trial that includes 300 credits, then $49.99/month unless canceled. With your own model API key, LLM usage costs no platform credits.
Can Versabot prove the results we deliver?
It can show a traceable chain from a verified receipt back to a conversion, lead and tracked click, and it excludes anything it cannot verify. It does not claim causal lift or automatically prove every offline or offsite conversion. Approved performance fees are invoiced through your own Stripe account, so money goes straight from your client to you.
Put the agent to work on one client first.
Connect what you already use, ask for the work in plain language, and approve what leaves the building.
7-day trial with a card on file and 300 usage credits. Then $49.99/month unless you cancel.