Kylon Pricing Explained: How Workspace Credits Actually Work
Kylon bills per workspace, not per seat. Here's how its credit system, agent tiers, and bring-your-own-subscription model affect what you pay.
What is Kylon and how does its pricing model work?
Kylon is a multiplayer AI workspace that looks like a cross between Slack and ClickUp, except the people in your channels can include AI agents alongside human teammates. Instead of charging per seat like most collaboration software, Kylon bills at the workspace level. One pool of credits covers everyone and everything inside that workspace, whether it’s a person typing a message or an agent running an automation in the background.
That distinction matters because the usual SaaS math (more people equals more cost) doesn’t directly apply here. What drives your bill in Kylon is usage, not headcount.
TL;DR
- Kylon charges per workspace rather than per member, so adding teammates to a workspace doesn’t automatically multiply your subscription cost.
- Usage is tracked through a shared credit pool, and the workspace owner can see how much each member or agent is consuming.
- Agents come in different intelligence tiers, commonly described as fast, pro, and max, and heavier tiers burn through credits faster.
- You can bring your own Claude Code or Codex subscription into a workspace, which lets an agent run on your existing plan instead of only consuming Kylon credits.
- Each workspace starts with a default agent (referred to in the product as a “head of agents”) that onboards you and can spin up additional agents, apps, and automations.
- Costs scale with how much work you ask agents to do: automations running on a schedule, background research, and app-building all draw down the same credit pool as live chats.
- Because billing is workspace-wide, teams are encouraged to treat each workspace like a business unit, isolating cost and activity for a specific project or department.
How does the per-workspace billing structure actually work?
When you create a new workspace, you’re prompted to name your team and optionally invite other members right away. Unlike tools that charge a flat fee per invited user, Kylon treats invitations as free. The workspace itself holds the credits, and every action taken inside it, whether by a human or an agent, draws from that same balance.
This is a meaningful departure from typical per-seat SaaS pricing. In most tools, inviting a tenth teammate costs the same as inviting the first. In Kylon, the question isn’t “how many people do we have” but “how much are those people (and their agents) actually doing.” On a team plan, workspace owners can view usage broken down by member, so it’s possible to see whose activity is consuming the most credits, whether that’s a human asking for constant revisions or an agent running scheduled automations every few hours.
What are the fast, pro, and max agent tiers?
Inside a Kylon workspace, agents aren’t a single flat capability. There are different intelligence tiers available, generally described as fast, pro, and max. The tiering reflects a tradeoff familiar to anyone who has used multiple models: lighter, faster agents cost fewer credits per action but may be better suited to routine tasks, while higher-capability tiers cost more per use and are reserved for harder reasoning or more complex builds.
This tiering gives teams a lever to control spend. A workspace that mostly needs an agent to check statuses, update a dashboard, or draft simple social posts can lean on a faster, cheaper tier. A workspace doing heavier lifting, like building out a full app or reasoning through a multi-step business process, can switch to a pro or max tier for that specific task without permanently paying max-tier rates for everything.
How do hosted agents differ from bring-your-own agents?
One of the more unusual aspects of Kylon’s billing is that it doesn’t force everything through its own credit meter. Workspaces can connect external coding subscriptions, specifically Claude Code and Codex subscriptions, directly into the environment. Support for additional providers such as Gemini, Hermes, and Grok is also planned.
When an agent runs on a connected subscription like this, it’s using that subscription’s own usage allowance rather than purely consuming Kylon credits. In practice, this means a workspace can have a head agent (Kylon’s native, credit-based agent) working alongside a “bring your own” Claude Code or Codex instance that’s billed separately through its own subscription. For teams that already pay for Claude Code or Codex and use them heavily, this hybrid setup can reduce how fast workspace credits get consumed, because some of the computational work is happening on a subscription the team is already paying for outside of Kylon.
Built like a system. Not vibe-coded.
Remy manages the project — every layer architected, not stitched together at the last second.
This is also the detail that addresses a common objection: why pay for another AI tool when you already have a coding subscription with a longer runway. The answer, structurally, is that Kylon isn’t necessarily replacing that subscription. It’s providing the collaborative layer (channels, shared memory, apps, automations) on top of agents that can still run on the subscriptions you already own.
What actually consumes credits in a workspace?
Credits aren’t just spent on chat messages. Several categories of activity pull from the same pool:
Conversations with agents in channels or direct messages are the most obvious draw. Beyond that, automations that run on a schedule (for example, an agent that checks news and rebuilds content every six hours) consume credits every time they execute, regardless of whether a human is watching. Building and iterating on in-workspace apps, such as dashboards or kanban-style boards, also uses agent time and therefore credits. Agent-to-agent collaboration counts too: agents proactively messaging each other to finish a task, without a human in the loop, still runs on the workspace’s credit balance.
This is worth factoring into any cost estimate. A workspace that only sees occasional chat activity will burn credits slowly. A workspace running multiple scheduled automations around the clock, on top of active app development, will consume credits considerably faster, even with the same number of human members.
Is Kylon’s pricing model worth it compared to per-seat tools?
Whether the model makes sense depends on how a team plans to use it. For teams already running multiple separate AI subscriptions per person (one Claude Code plan each, one Codex plan each, disconnected skills and automations per individual), consolidating into a shared workspace can reduce duplicated effort. Instead of five people each building their own integration to the same Google Drive or ClickUp instance, one workspace can host shared apps and automations that everyone’s agents read from and write to.
The workspace-based credit pool also means cost scales with actual output rather than team size, which can favor smaller, high-usage teams over larger teams with light AI usage. The tradeoff is that heavy automation use, especially multiple agents running on scheduled intervals across several channels, can consume credits quickly, so teams need to monitor usage much as they would monitor cloud compute spend rather than treating it like a flat monthly seat license.
Because Kylon lets workspaces bring in outside subscriptions, teams that already have Claude Code or Codex access can offset some of that cost by routing heavier agent work through those existing subscriptions rather than relying solely on Kylon’s native credits.
Frequently Asked Questions
Does Kylon charge based on how many people are in a workspace?
No. Kylon’s billing is tied to the workspace itself, not the number of members in it. Inviting more human teammates doesn’t directly increase the subscription cost; what affects cost is how much activity (chats, automations, app building) happens inside the workspace.
What are the fast, pro, and max agent tiers in Kylon?
They represent different levels of agent capability and cost. Faster, lighter tiers are cheaper per action and suited to routine tasks, while pro and max tiers cost more per use and are intended for more complex reasoning or development work.
Can I use my existing Claude Code or Codex subscription inside Kylon instead of paying for everything through credits?
Remy doesn't build the plumbing. It inherits it.
Other agents wire up auth, databases, models, and integrations from scratch every time you ask them to build something.
Remy ships with all of it from MindStudio — so every cycle goes into the app you actually want.
Yes. Kylon allows workspaces to connect external subscriptions, including Claude Code and Codex, so agents can run on those existing plans. This can reduce how much a workspace relies on its own native credit pool for heavier workloads.
What consumes credits fastest in a Kylon workspace?
Scheduled automations that run repeatedly (for example, every few hours), active app development, and ongoing agent-to-agent collaboration tend to consume credits faster than occasional one-off chat messages.
Is Kylon cheaper than running separate AI subscriptions for every team member?
It depends on usage patterns. Teams that were previously paying for individual subscriptions per person and duplicating integrations may find a shared workspace model more efficient, but teams running heavy, constant automation across many channels should expect credit consumption to scale with that activity.
