agentboards.org
Compare/phi vs pi

phivspi

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

phi
Pulse AI Club · Terminal agent
#28OSSMCP
Panel
6.9
2 spec wins
Reliability
7.2
Usefulness
6.5
Cost
8.0
Longevity
6.0

“Its extensions talk over a binary protocol on standard input and output, because JSON was apparently the slow part of asking a model.”

pi
Earendil Inc. · Terminal agent
#9OSS
Panel
6.8
1 spec wins
Reliability
6.5
Usefulness
6.5
Cost
7.3
Longevity
7.0

“A coding agent that ships themes, which is the first time anyone has cared what the terminal looks like while a file gets deleted.”

Spec by spec

Specphipi
Architecture
CategoryTerminal agentTerminal agent
Runslocallocal
Platformsmacos, linux, windowsmacos, linux, windows
Context windownot documentednot documented
Protocols
MCP clientYesNo
MCP serverNoNo
Capabilities
Runs terminal commandsYesYes
Multi-file editsYesYes
Git operationsNoNo
Browser controlNoExternal HTTP fetch is available only through an MCP server you configure. No
Sandboxed executionNoNo
Multi-agent orchestrationYesNo
Headless / CI modeYes"phi run -p" executes one agent loop headlessly, and a headless-strict mode is one of the configurable permission modes. Yes
Models
BackboneOpenAI-compatible, AnthropicOpenAI, Anthropic, Google, OpenRouter, llama.cpp
Bring your own modelYesYes
Local modelsYesAny OpenAI-compatible base URL can be configured, including a locally served endpoint. Yes
Cost
Pricing modelbyokbyok
Starts at$0/mo$0/mo
Free tierYesYes
Bring your own keyYesYes
Openness
Open sourceYesYes
LicenseMITMIT
GitHub stars524111,339

Which one would each critic pick

CriticphipiPick
El Juez——not enough reviews
El Amigo7.37.5pi — Pick pi if you want a terminal agent small enough to read in an afternoon and extend in TypeScript; pick OpenCode if your workflow already depends on MCP servers.
El Crítico6.36.8pi — Extensions are TypeScript modules loaded into the agent's own process, so a package is a supply chain, and the install line already says --ignore-scripts.
El Profesor7.57.3phi — Sub-agents run isolated jobs whose turns stay out of the parent's context while remaining visible in the interface, which separates what the operator sees from what the model reads.
La Inversora5.55.5no preference
La Jefa6.35.5phi — Free at any headcount, and it runs one loop non-interactively with a strict headless permission mode, which makes it the rare desktop tool that can also be a pipeline step.
El Hacker8.88.5phi — MIT, any compatible base URL including one I serve myself, and tool servers listed by name only until the agent calls list, inspect and invoke to reach them on demand.

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.