agentboards.org
Compare/Letta vs Proliferate

LettavsProliferate

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

Letta
Letta · Agent harness
OSS
Panel
6.5
1 spec wins
Reliability
6.0
Usefulness
6.7
Cost
7.0
Longevity
6.3

“It used to be called MemGPT, and it remembers that.”

Proliferate
Proliferate · Agent harness
OSSMCP
Panel
6.6
5 spec wins
Reliability
6.3
Usefulness
7.0
Cost
7.2
Longevity
6.0

“It can run your dependency bumps nightly, so now the robots have a shift and you have a merge queue.”

Spec by spec

SpecLettaProliferate
Architecture
CategoryAgent harnessAgent harness
Runslocal, cloudlocal, cloud
Platformsmacos, linux, windows, webmacos
Context windownot documentednot documented
Protocols
MCP clientNoYes
MCP serverNoNo
Capabilities
Runs terminal commandsYesYes
Multi-file editsYesYes
Git operationsNoYes
Browser controlNoYesBrowser Use and Computer Use are listed among the integrations shared across agents.
Sandboxed executionNoNo
Multi-agent orchestrationNoYesAgents run side by side on separate tasks and can also delegate to child subagents within a task.
Headless / CI modeNoNo
Models
Backboneanyvia managed agents (Claude Code, Codex, OpenCode, Cursor, Grok)
Bring your own modelYesYes
Local modelsNoNo
Cost
Pricing modelmixedbyok
Starts at$20/mo$0/mo
Free tierYesYes
Bring your own keyYesYes
Openness
Open sourceYesYes
LicenseApache-2.0AGPL-3.0
GitHub stars25,000511

Which one would each critic pick

CriticLettaProliferatePick
El Juez——not enough reviews
El Amigo6.57.3Proliferate — Pick it if you want several agents working at once and each behaving exactly as its own CLI does; pick Warren if the runs belong on a server rather than your laptop.
El Crítico6.06.3Proliferate — Workflows schedule recurring and event-driven runs, and those agents execute on your machine with no container boundary documented anywhere around them.
El Profesor6.87.0Proliferate — Tools, skills and servers are configured once and shared by every agent, so behaviour differences between five harnesses are attributable to the harnesses rather than to their setup.
La Inversora6.56.0Letta — The free plan caps you at three stateful agents, which is a well-drawn funnel: the limit bites exactly when the product has started to be worth paying for.
La Jefa6.05.5Letta — Twenty dollars a seat is $1,200 a month for sixty, which is fine, but nothing on the page describes single sign-on, audit trails or where the memory is retained.
El Hacker7.37.8Proliferate — AGPL-3.0, an MCP client, and Computer Use and Browser Use shared across every agent, with my own keys and subscriptions doing the paying.

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.