MCP Index

Agent-Ready Software

Software that's actually easy for AI agents to use — not just software that mentions AI.

Browse Tools
Browse all tools →

Agent-ready tools (25)

Source: anyCan do: anythingAuth: anyScore: anySort: Best overall
Agent-ready tools
ToolCost to actCan doRisk
GitHubExcellentDeveloper Tools · The category default: official, enormous tool surface, both local and hosted. Watch which path you take — the hosted one drags Copilot licensing along.
5
5
5
4
5
FreeUnmeteredAdminHeadlessHighblast radius
DigitalOceanExcellentCloud Infrastructure · Per-service tool docs published in llms.txt and a simple PAT instead of IAM — far easier than AWS for the same class of work.
5
5
5
4
5
FreeUnmeteredAdminHeadlessHighblast radius
Jira (Atlassian)ExcellentProject Management · Widest product coverage of any PM server; the gate is rate quota by plan tier, not access.
4.5
4.5
5
4.5
5
FreeUnmeteredCore writeHeadlessMediumblast radius
SentryExcellentObservability · The reference example of a debugging-loop server: official, remote, free tier, and it pairs with sentry-cli for release work.
4.5
4.5
5
4.5
5
FreeUnmeteredCore writeHeadlessMediumblast radius
GitLabExcellentDeveloper Tools · Two good paths — hosted server or glab CLI in MCP mode — and self-managed instances are supported, which GitHub's hosted endpoint can't match.
4.5
5
5
4
4.5
FreeUnmeteredAdminHeadlessHighblast radius
NeonExcellentDatabase · Best agent-safety story in databases: instant branches mean a destructive agent action is a discarded branch, not an incident.
4.5
5
5
3.5
5
FreeUnmeteredAdminNeeds a humanLowblast radius
StripeExcellentPayments · The most complete agent story in payments: official server, restricted keys, a real CLI and 92KB of llms.txt. Write tools move real money.
4.5
4.5
5
4
5
FreeUnmeteredCore writeHeadlessHighblast radius
WebflowExcellentDesign · Strongest agent story among site builders: official, open-source, OAuth, and element/component/style writes no longer need a human in the Designer.
4.5
4.5
5
4
5
FreeUnmeteredCore writeHeadlessMediumblast radius
MongoDBExcellentDatabase · Covers both data and infrastructure, which is unusual — and means an agent can spin up billable clusters. Scope the API key.
4.5
4.5
5
4
5
FreeUnmeteredAdminHeadlessHighblast radius
SupabaseExcellentDatabase · Among the most capable database servers — which is the risk: unrestricted, an agent can alter schema. Read-only flag first, loosen deliberately.
4.5
4.5
5
4
5
FreeUnmeteredAdminHeadlessHighblast radius
NetlifyExcellentCloud Infrastructure · Write-capable where Vercel is not — an agent can actually ship a site and wire up functions end to end.
4.5
4.5
5
4
5
FreeUnmeteredAdminHeadlessHighblast radius
GrafanaExcellentObservability · Free, open-source and write-capable on dashboards — the default pick if you self-host your observability stack.
4.5
4
5
4.5
5
FreeUnmeteredCore writeHeadlessMediumblast radius
monday.comExcellentProject Management · Was among the first five official servers in the wild; read/write is solid, watch the per-plan API complexity budget.
4.5
4.5
5
4
4.5
FreeUnmeteredCore writeHeadlessMediumblast radius
PagerDutyExcellentObservability · Both hosted and local options, plus a published MCP OpenAPI spec — one of the more thoughtfully shipped on-call servers. Write scopes touch live incidents, so gate them.
4.5
4.5
5
4
4.5
FreeUnmeteredCore writeHeadlessMediumblast radius
CalendlyExcellentEmail & Calendar · One of the few schedulers that chose a hosted remote server — no local process, standard OAuth, and it works before you pay.
4.5
4.5
5
4
4.5
FreeUnmeteredCore writeHeadlessMediumblast radius
ZoomExcellentCommunication · Official MCP docs exist and the API underneath is mature; the practical gate is that the useful data (recordings, transcripts) lives on paid plans.
4.5
4.5
5
4
4.5
FreeUnmeteredCore writeHeadlessMediumblast radius
RampExcellentBanking & Finance · The most complete finance-ops surface in this category, and scope-gated rather than all-or-nothing. Grant read scopes first.
4.5
4.5
5
4
4.5
FreeUnmeteredCore writeHeadlessHighblast radius
VercelExcellentCloud Infrastructure · Great for debugging a deployment, useless for changing one — the MCP server is read-only, so writes still go through the CLI.
4.5
3.5
5
5
5
FreeUnmeteredRead onlyHeadlessLowblast radius
RailwayExcellentCloud Infrastructure · Fastest path from 'agent has a repo' to 'agent has a running deployment' — local and hosted servers, GraphQL API, solid CLI.
4.5
5
4
4
4.5
FreeUsage-meteredAdminHeadlessHighblast radius
CloudflareExcellentCloud Infrastructure · Highest agent-readiness in infrastructure: multiple official servers, wrangler CLI, OAuth, free tier, and first-class docs for agents.
4.5
5
5
2.5
5
FreeUnmeteredAdminNeeds a humanHighblast radius
AttioExcellentCRM · Built after MCP existed and it shows — hosted server, clean object model, write access without an enterprise conversation.
4.5
4.5
5
3.5
4.5
FreeUnmeteredCore writeHeadlessHighblast radius
NotionExcellentProject Management · Strong read/write coverage on free plans; the paywall is on AI search and admin controls, not on MCP itself. No bearer tokens — interactive OAuth only.
4.5
4.5
5
3
5
FreeUnmeteredCore writeNeeds a humanMediumblast radius
PostHogExcellentProduct Analytics · Best-in-class: official server, real write actions (flags, insights), free tier, and docs published as llms.txt.
4.5
4.5
5
3
5
FreeUnmeteredCore writeHeadlessMediumblast radius
PrismaExcellentDatabase · Unusually well-matched to coding agents: the MCP server, the CLI and the schema language all target the same workflow.
4.5
4.5
5
3
5
FreeUnmeteredAdminHeadlessMediumblast radius
ReachCostControlTrust

What makes software "agent-ready"?

Agent-ready software gives AI agents a reliable way to read and act on its data: an official MCP server, a documented public API, a CLI, an llms.txt file, and clear, agent-friendly documentation. The bar isn't whether a product has an AI feature — it's whether an agent can actually integrate with it.

The real question

The useful question isn't "does this tool have MCP?" It's "can my agent actually do useful work with this tool?" That's why every listing here includes an agent-readiness score, not just a badge.

How we score it

  • MCP status: official, community, none, or unknown.
  • CLI and public API availability.
  • llms.txt and agent-specific documentation.
  • Read and write action support.
  • Authentication difficulty (none, API key, OAuth, complex).
  • Verified compatibility with Claude, Cursor, ChatGPT, Windsurf, and other agents.

Related