this+that vs Drag
A Gmail queue your team works together, or every channel read as incoming work
Drag turns Gmail into a shared inbox for customer email, with boards, assignment and an AI layer that drafts and resolves replies, and its MCP server lets Claude and other AI clients run that inbox by prompt. this+that starts somewhere else. It reads Gmail, Outlook, Slack, Microsoft Teams, Google Chat, and WhatsApp Business as one stream of incoming work, then pulls the requests, deadlines and commitments out of those messages into tracked tasks.
A customer mailbox a team shares, or every channel a company runs on
Drag turns Gmail into a shared inbox for customer email. A team connects support@ or sales@, works the queue together on kanban or list boards, assigns threads, leaves internal notes and gets a warning before two people answer the same message. The mail stays in Gmail, so adopting it means an extension and a login rather than a migration. On top of that sits Drag AI: Copilot answers questions about the inbox, six Assists draft and classify inside a thread, Intelligence grounds those answers in your knowledge base, and Agents resolve routine email end to end with every source cited.
Their headline is "The inbox for teams and agents", which is close to where we think email is heading, and Drag reached the agent half of it early. Their MCP server hands Claude, ChatGPT, Gemini, Copilot and Cursor 47 tools across 12 categories, read and write, covering email, boards, contacts, the knowledge base, WhatsApp and analytics. It is open source under MIT, connects from npm or a hosted URL, and inherits whatever permissions the person signing in already has. Our own MCP server exposes one read tool for DoBox today, with writes on the roadmap, so on that comparison Drag is ahead of us.
Where the two products diverge is what the inbox is for. Drag is purpose-built for customer email: one mailbox and one team working it together, with a WhatsApp channel and an early-access live chat alongside. this+that reads Gmail, Outlook, Slack, Microsoft Teams, Google Chat, and WhatsApp Business for one company, and treats an arriving message as work to be done rather than a conversation to be answered. A request buried in a long Slack thread and a new billing address sitting in an email signature both count, and they come out as tasks in DoBox, with contacts that keep themselves current from the same messages and a brain that workflows write back to as they run. Working a support queue inside Gmail together is what Drag is built for, and we are not trying to be. Use us when the work reaches you across every channel the company runs on.
Feature by feature
How the two products compare across the things that matter most.
| Topic | Drag | this+that |
|---|---|---|
| Core use case | Customer email, worked by a team in one place. Boards, assignment, internal notes, collision detection and a help center. | A work layer over every connected channel. Arriving messages become tracked tasks, contacts and knowledge that workflows act on. |
| Where it runs | Inside Gmail and Google Workspace, as a Chrome extension, plus web, desktop, iOS and Android apps. Your email stays in Gmail. | A web app that rebuilds the inbox across every connected channel, plus a Chrome extension that puts DoBox inside Gmail. |
| Channels | Email at the core, with a WhatsApp channel on Pro and live chat in early access. | Gmail, Outlook, Slack, Microsoft Teams, Google Chat, and WhatsApp Business, read as one stream. |
| Unit of work | The conversation. A thread becomes a card that moves across columns and belongs to a person. | The task inside the conversation. AI reads every message and lifts out requests, deadlines and commitments before anyone opens the thread. |
| Automation | Rules that tag, route and auto-respond. Business hours, SLA rules and round-robin assignment come with Pro. | Plain-language workflows fired by an arriving message, with conditional logic, loops, MCP tool calls and sub-agents. |
| AI in the product | Copilot for questions about the inbox, six Assists inside a thread, Intelligence for grounding, and Agents that resolve routine email. Included on Plus and Pro, with Agent usage billed separately. | AI reads every channel to extract the work, drafts replies grounded in the brain, and runs as steps inside workflows. |
| MCP | 47 tools across 12 categories, read and write, open source under MIT, for Claude, ChatGPT, Gemini, Copilot and Cursor. On the Pro plan. | MCP in both directions. Workflows call any MCP server as a step, and our own server exposes one read tool for DoBox today. |
| Knowledge | A help center and knowledge base that ground replies and cite their sources, with articles readable and writable over MCP. | The brain: operational knowledge with version history that grounds drafts, answers and workflow steps, and that workflows write back to as they run. |
| Contacts | Shared contacts inside the inbox with their conversation history, on Plus and Pro. | An address book that assembles itself from the channels you connected, with the correspondence attached. |
| Reporting | Response times, workload and SLA breaches, over a window that widens with the tier. A strength for anyone running a queue. | Not a focus. Work is tracked as tasks rather than as queue-level response-time metrics. |
| Pricing | Starter is $12 per user per month billed yearly, or $16 billed monthly. Plus adds the AI layer at $18 and $22. Pro, which carries the API, MCP access and the WhatsApp channel, is $24 and $30. A seven-day trial runs with full Pro access and no card. | Free for 30 days. |
| Scale and compliance | Drag reports more than 200,000 users and a 4.7 rating on G2 from over 1,200 reviews, and lists CASA and SOC 2 certification. | Out of beta, with SOC 2 Type I as of July 31, 2026 and Type II underway, alongside CASA Tier 2. |
- Email never leaves Gmail. Adopting Drag costs an extension and a login, not a migration.
- Their agents cite sources and log every action.
- Their MCP server is the most complete we have seen in this category: 47 tools across 12 categories, read and write, MIT licensed, reachable from npm or a hosted URL, inheriting the permissions of whoever signed in.
- Working a queue together is the whole product. Boards, assignment, collision detection, internal notes, round-robin.
- Reporting on response times, workload and SLA breaches is mature.
- Drag publishes more than 200,000 users and a 4.7 rating on G2 across 1,200-plus reviews.
- The AI layer is inside the plan, not a per-seat add-on. The bill is predictable.
- We read Gmail, Outlook, Slack, Microsoft Teams, Google Chat, and WhatsApp Business, so a request that arrives outside the customer mailbox is caught with everything else.
- The task is the unit. Requests, deadlines and commitments are lifted out of messages into DoBox and assigned there, instead of the thread itself moving across a board.
- Contacts build themselves from those same channels, correspondence attached.
- Workflows fire on an arriving message. They branch, loop, call any MCP server and hand jobs to sub-agents, so routine work finishes without anyone opening it.
- The brain grounds drafts, answers and workflow steps, and keeps version history. Workflows write back to it as they run.
- Nothing has to be a support queue first. A founder or an ops lead can point this+that at the channels they already live in without standing up a team mailbox.
See what's waiting in your inbox
Connect your inbox and this+that starts reading it. The tasks, follow-ups, and open commitments buried in your messages become tracked work, and what your team knows starts collecting in one place.
Free for 30 days · No credit card required