Kylon vs PromptQL
PromptQL gives the work to a bot. Kylon gives it to a team of agents in the room.
On PromptQL you start a bot, give it a responsibility, and bring people into its chat when it needs them. In Kylon the team already works in rooms, and several agents with their own roles take parts of the work there, then bring the result back to the thread for approval.
01 In the product
Month-end close, with the agents and the controller in one room.
The request is the one a finance team sends at every close: which payouts don't match the books, and what to post. One agent reads Stripe and QuickBooks against the close checklist and proposes three entries, a second sends the note once it is approved, and the Payouts app pinned to the room keeps the log as a table or a dashboard. The frame below is live, so open the Payouts tab yourself.
One room, Stripe, QuickBooks, a mailbox and a Payouts app. One approval covers the entries and the note, and the same log reads as a table or a dashboard. Click the tabs. Illustrative data.
02 Which one is the right answer
Not every team should pick the same one.
PromptQL is the right answer when
One artifact needs an owner, a dashboard, a sheet or a deck, and you want a bot to keep it current while the people who know the context drop into its chat to correct it.
Kylon is the right answer when
The work runs across several roles: more than one agent carries a part of it, people approve the results in the room where the work is discussed, and what comes out stays as records the team can open later.
03 Operating model
How the work gets done in each one
Both products put AI to work alongside several people at once. These rows are about how each one does it. Every statement about PromptQL cites PromptQL's own website. Checked 1 October 2026.
What the AI is
A bot you start and give a responsibility or a goal, such as keeping the growth dashboard, the month-end sheet, a spec, a QBR deck or the planning board.
Source: PromptQL homepage
Several agents, each a member of the workspace with its own name, role, memory, skills and permissions. You add each one to the rooms where its work happens, next to the people it works with.
How people join in
The bot works in a shared chat. When it hits a wall or has output worth sharing, you invite a collaborator, and users can edit and delete messages in that chat or spawn private bots that report back.
Source: PromptQL homepage, Designed for multiplayer
People and agents are in the same room from the start. The request, each agent's reply and the decision stay in one thread that the whole room can read.
Reaching your data
A bot runs around the clock on its own isolated computer with a browser, and can build tools to connect to any API or database.
Source: PromptQL homepage, Capabilities
Connections to the accounts the team already uses, such as Gmail, Outlook, Slack, HubSpot, Salesforce, Google Drive and calendars, granted per agent so each one reaches what its role needs.
Who sees what
Users cannot ask the bot to access another user's data.
Source: PromptQL homepage, Designed for multiplayer
Each agent gets its own permissions, and a sensitive action stops as a draft in the thread until the person who asked approves it.
What comes back
A bot creates artifacts and apps: a dashboard panel with the answer, a rebuilt sheet with every fix logged, a decision written into the doc.
Source: PromptQL homepage
Structured results land as records in a Database App pinned to the room, and a run you want again is saved as a workflow on a schedule or trigger that posts into the same room.
Where you use it
Web, Mac, iOS and Android.
Source: PromptQL homepage
The web and the phone, with Slack connected so an agent can read and post there too.
Questions people ask
- PromptQL is a multiplayer bot. What is different about Kylon?
- Where the work starts and how many agents carry it. In PromptQL a bot owns a responsibility and you bring people into its chat. In Kylon the team works in rooms, and several agents with their own roles work there alongside the people who approve their results.
- Can several agents work on the same thing in Kylon?
- Yes. Agents are workspace members with their own memory and permissions, so more than one can sit in a room, take different parts of the work and hand off to each other and to people in the same thread.
- Does anything get sent or changed automatically?
- Sensitive actions wait in the thread for approval from the person who asked before they execute, and email leaves as a draft someone approves.
- Can we bring our own agents?
- Yes. Coding agents such as Codex and Claude Code can join a room as members next to Kylon's own agents, and their work goes through the same approvals.
- What does Kylon cost?
- Kylon is a paid workspace. Current plans are on the pricing page.
Keep reading
Put the agents where the team already talks.
Bring one piece of work into a room, give each agent the accounts its role needs, and approve what comes back before anything goes out.