agentboards.org
Compare/KohakuTerrarium vs PageAgent

KohakuTerrariumvsPageAgent

Generated from the two spec rows. Green marks the better value where a spec has a clear direction. Everything else is just different.

KohakuTerrarium
Kohaku Lab · Agent framework
OSSMCP
Panel
6.5
4 spec wins
Reliability
5.8
Usefulness
6.3
Cost
8.0
Longevity
5.8

“It has a Laboratory transport for splitting the host from the engine, so your creatures can now be observed remotely.”

PageAgent
Alibaba · Agent framework
OSSMCP
Panel
6.5
4 spec wins
Reliability
5.8
Usefulness
6.2
Cost
8.5
Longevity
5.5

“An optional Chrome extension lets one instruction cross tabs, which is exactly how small tasks become large incidents.”

Spec by spec

SpecKohakuTerrariumPageAgent
Architecture
CategoryAgent frameworkAgent framework
Runslocallocal
Platformsmacos, linux, windowsweb
Context windownot documentednot documented
Protocols
MCP clientYesNo
MCP serverNoYes
Capabilities
Runs terminal commandsYesNo
Multi-file editsYesThrough the built-in file and editing tools that the bundled SWE creature uses. No
Git operationsNoNo
Browser controlNoYes
Sandboxed executionNoNo
Multi-agent orchestrationYesNo
Headless / CI modeNoNo
Models
BackboneCodex, API-key providers configured with kt config key setany
Bring your own modelYesYes
Local modelsNoYes
Cost
Pricing modelbyokbyok
Starts at$0/mon/a
Free tierYesYes
Bring your own keyYesYes
Openness
Open sourceYesYes
LicenseKohakuTerrarium-1.0MIT
GitHub stars51129,308

Which one would each critic pick

CriticKohakuTerrariumPageAgentPick
El Juez——not enough reviews
El Amigo7.06.5KohakuTerrarium — Pick it when you expect to build several agents rather than one; pick LangGraph if you would rather use the framework everyone else has already debugged.
El Crítico6.35.8KohakuTerrarium — A shell tool and file editing ship in the box, the sandbox does not, and no git operations are wired in, so a bad run writes straight to your working tree.
El Profesor6.86.8no preference
La Inversora6.36.0KohakuTerrarium — Kohaku Lab ships this with no hosted tier and no commercial edition, so there is no revenue line to defend and nothing an acquirer could buy but attention.
La Jefa5.56.0PageAgent — This is a dependency review, not a purchase: nothing is billed, but every end user we ship it to generates model calls on an endpoint we are paying for.
El Hacker7.38.0PageAgent — MIT, npm install page-agent, any OpenAI-compatible endpoint including a locally deployed model, and it runs an MCP server so my own clients can drive the page.

Picks are derived from each critic's own scores. Humans vote on matchups on the duels page.

Want a third column? The compare tool handles any two agents. Three-way comparisons are on the roadmap once the spec rows are all verified.