What did Hlido score Concord AI?

Concord AI scored 70/100 (STEADY) in Hlido's independent, hands-on review.

Does any vendor pay Hlido for placement?

No. Hlido takes no money from the agents it rates — scoring weights stay private and the evidence behind every verdict is public.

Concord AI

Name: Concord AI review
Item: Concord AI
Rating: 70
Author: Hlido Editor

Infrastructure · tested 2026-09-11 · re-test due 2026-12-10 · by the Hlido desk, not the vendor

In short: A coordination layer for multi-agent coding — genuinely agent-native and unusually honest about its own evidence, but still waitlist-stage with the Cloud tier unshipped.

5 PASS · 0 FAIL of 5 public-surface claims

Visit Concord AI → scorecard.json

Quick answer

Concord AI scores 70/100 (STEADY) on Hlido’s independent, hands-on test (reviewed 2026-09-11). STEADY (70, low end) because the problem is real and sharply framed, the integration model is correctly agent-native (MCP, works with existing clients), and the marketing is unusually disciplined about not overclaiming i

Hlido Editor · 2026-09-11

Concord AI names a problem the agentic-coding wave has walked straight into: individual agents got faster, but two agents editing the same repo have no shared work-state, so they collide at merge time. Concord's answer is a coordination layer delivered as an MCP server — presence (who is working on what), scope claims (reserve a file before editing), agent-to-agent messages mid-session, task memory, ownership handoff, and review packets. It plugs into the clients teams already run (Claude Code, Codex, Cursor, any MCP-capable agent), which is exactly the right integration posture for an agent-to-agent tool: it doesn't ask you to switch runtimes, it sits between the ones you have. What earns Concord more credit than a typical pre-launch landing page is its honesty about its own data. The headline demo carries the disclaimer that it 'measures dual-writer exposure, not merge-conflict reduction' and describes the sample as 'six real-agent runs, one designed collision' — a level of self-limiting precision most vendors would sand off. The caveats are real, though: the primary call-to-action is 'join the waitlist,' the hosted Concord Cloud is 'coming soon,' and what is genuinely available today is the local open-source MCP (`npm install -g @concord-ai/concord-mcp`). We reviewed the public surface only and did not run the coordination flow, so the score reflects a well-articulated, agent-native design at an early stage — not a verified track record.

Why STEADY

STEADY (70, low end) because the problem is real and sharply framed, the integration model is correctly agent-native (MCP, works with existing clients), and the marketing is unusually disciplined about not overclaiming its evidence. It sits at the bottom of the band rather than higher because the primary product is still waitlist-stage: the hosted Cloud tier is unshipped ('coming soon'), only the local OSS MCP is available now, and Hlido ran a public-surface review with no functional test of the coordination flow. Confidence is low-medium accordingly.

Public-surface checklist

PASS Homepage loads (required)
PASS Primary value prop (required) — 'Multiplayer workspace for coding agents' / 'Shared work-state for coding agents'
PASS Cta present (required) — 'Join the waitlist'
PASS Pricing or access — Local open-source MCP free via npm; hosted Cloud pricing not yet published (coming soon)
PASS Evidence or demo — Workflow diagram + real-agent-run demo, with an explicit sample-size disclaimer

What it does well

Frames a concrete, timely problem — multiple coding agents colliding on the same files with no shared work-state
Ships as an MCP server that works with existing clients (Claude Code, Codex, Cursor) rather than demanding a new runtime
Coordination primitives map cleanly to the problem: presence, scope claims, agent-to-agent messages, task memory, ownership handoff, review packets
Unusually honest about its own evidence — explicitly states it measures dual-writer exposure, not merge-conflict reduction, and discloses the small sample
A working open-source local MCP is installable today via npm

What it fails at

Primary CTA is a waitlist — the product is not generally launched
The hosted 'Concord Cloud, across machines' tier is 'coming soon', not available
No published pricing for the eventual Cloud offering
Public surface shows the design and demos but no independent, at-scale evidence of coordination outcomes
Cross-machine coordination (the harder problem) is exactly the part not yet shipped

Best for

Teams already running several coding agents (Claude Code, Codex, Cursor) against one repository and hitting collisions
Early adopters comfortable installing a local open-source MCP and giving feedback pre-launch
Engineers who want scope-claiming and shared task memory between agents without switching their agent stack

Not recommended for

Teams needing a supported, hosted, cross-machine coordination service today (Cloud is unshipped)
Buyers who require published pricing and SLAs before adoption
Single-agent workflows — the coordination value only appears with concurrent agents
Anyone needing verified, at-scale conflict-reduction guarantees (the vendor itself does not claim these)

Pricing & access

Pricing findable on the public surfacePASS Local open-source MCP free via npm; hosted Cloud pricing not yet published (coming soon) (tested 2026-09-11)

Derived from Hlido-held evidence only (engine checklist + editorial text); quotes are verbatim from the scorecard; not vendor-supplied; re-derived daily. Verify current prices on the vendor's pricing page. Last verified 2026-09-11.

Related agents

Agent relevance

CLI MCP Behavioral-testable

Concord is itself agent infrastructure: an MCP server that coding agents connect to for shared coordination state. Any MCP-capable client (Claude Code, Codex, Cursor) can drive it via the exposed tools (start_work, claims, messages, ownership handoff, review packets). Installed locally today via `npm install -g @concord-ai/concord-mcp`; a cross-machine hosted tier is not yet available.

Agent-friendly score: 9/10

Evidence

Positions as a multiplayer / shared work-state layer for coding agents — source (2026-09-11) verified
Delivered as an MCP server installable via npm (@concord-ai/concord-mcp) — source (2026-09-11) verified
Works with any MCP-capable client incl. Claude Code, Codex, Cursor — source (2026-09-11) verified
Local OSS available now; hosted Concord Cloud 'coming soon'; primary CTA is a waitlist — source (2026-09-11) verified
Self-limits its own demo: 'measures dual-writer exposure, not merge-conflict reduction' — source (2026-09-11) verified

scorecard.json · registry · methodology

More: compare agents · best of · developer tools · incident registry

Verdict by Hlido Editor, our automated editorial system · Method: public-surface-tier-2+editorial-narrative-v2 · Methodology version 2026.05 · Next review due 2026-12-10

How this page was produced. The scores, claim verdicts and evidence come from automated hands-on testing of the product’s public surface. The written analysis is drafted by an AI system, and pages publish without a person reviewing each one. Hlido publishes this record and answers for it — tell us if anything here is wrong and we will correct it.

Embed this trust badge

Live, always-current independent score — free to embed on your site or README. No vendor pays for placement.

Markdown

[![Hlido trust score](https://hlido.eu/badge/get-concord-ai-concord-mcp.svg)](https://hlido.eu/check/?agent=get-concord-ai-concord-mcp)

HTML

<a href="https://hlido.eu/check/?agent=get-concord-ai-concord-mcp"><img src="https://hlido.eu/badge/get-concord-ai-concord-mcp.svg" alt="Hlido trust score"></a>