Skip to main content
AGENTFRAMETry Agentframe
AGENTFRAME
FIND YOUR WAY
Try Agentframe

Give good ideas somewhere to go.

GROW AT YOUR OWN PACE

Small beginnings.
Room to grow.

Start with an idea. Find your rhythm. Scale when you’re ready.

A little less. −20%

Tinker

For the idea you can’t leave alone.

$0/ month

Free to explore. No card needed.

  • 1,000 agent runs / month
  • 3 projects to play with
  • Core tools & persistent memory
  • Community support
  • Not included: Shared team workspaces

Builder

THE SWEET SPOT

For the things you’re ready to ship.

$32/ month

$384 billed yearly

  • 25,000 agent runs / month
  • Unlimited projects
  • Everything in Tinker
  • 30-day execution history
  • Shared workspaces & approvals

Team

For good work, at a bigger scale.

$96/ month

$1152 billed yearly

  • 100,000 agent runs / month
  • Everything in Builder
  • 90-day execution history
  • Roles & workspace controls
  • Priority human support

Bring your own model keys. Model usage is billed by your provider.All plans include a place to make something worthwhile.

THE DETAILS MAKE THE DIFFERENCE

More than the
next response.

Useful context. Room to choose. A way to pick up where you left off. See what comes with your workspace.

Agentframe Builder, Runloom Launch, and Taskgrove Studio: a comparison of fictional paid plans.
At a glanceThe paid plansAgentframeBuilder planRunloomLaunch planTaskgroveStudio plan
01Context that carries forward
Workspace memoryAcross runsPer sessionAcross runs
Run recoverySaved checkpointsRestart the runManual checkpoints
02Tools on your terms
Model choiceBring your ownHosted modelsBring your own
MCP toolsNative supportVia an adapterNative support
Human approvalsPer toolPer workflowTeam upgrade
03A clear view of the work
Execution history30 days7 days14 days
Shared workspacesIncludedAvailable as an add-onIncluded

Agentframe, Runloom, and Taskgrove are fictional platforms. Plans and capabilities are illustrative.

EXPLORE THE HARNESS
A FEW GOOD QUESTIONS

Glad you asked.

A little more context,
before you get going.

EXPLORE THE EXAMPLES

Think of a harness as the working environment around a model. The model provides the intelligence; the harness connects it to tools, keeps its context, manages the execution loop, and gives people a way to inspect and guide the work.