Cleric

Cleric is an AI SRE agent that verifies production changes, investigates alerts and proposes fixes for engineering teams that run their own services. Its focus is the gap between merging code and knowing it works in production: Cleric keeps watching after each release and connects any regression back to the change that caused it. When it does page a human, the page comes with a cause and evidence rather than just a red graph.

What is Cleric used for?

Teams use Cleric to reduce noisy pages, speed up investigations and catch regressions that slip past tests. It reads from your existing tools and only writes when you allow it.

• Verifying releases by monitoring production for about 14 days after a pull request merges
• Taking on-call alerts, closing false positives and paging engineers only when needed
• Investigating issues to a root cause with supporting evidence
• Proposing a fix as a pull request and confirming the issue clears after deploy
• Answering questions in Slack, Microsoft Teams or the web app
• Giving coding agents production context through MCP

Who Cleric is best for

Cleric fits startups and scale-ups with real production traffic, an on-call rotation and enough alerts that triage eats engineering time. Platform and SRE teams that want to measure value per task will appreciate the credit model. Very small teams with few alerts may find the $700 monthly entry price hard to justify, and companies that cannot grant any external read access to production systems will struggle to use it.

Cleric pricing

As of October 2026, Cleric's Team plan costs $700 per month ($8,400 per year) with 1,400 credits a month; Pro costs $1,500 per month with 3,000 credits and adds scheduled agents, private resource access and SSO; Scale costs $3,000 per month with 6,000 credits and a shared Slack channel with Cleric engineers. Enterprise is annual only with a custom credit pool, data residency options and a named success manager. Credits cost $0.50 each: verifying a production change uses 2 credits, investigating an issue uses 10, and chats are free under 30 seconds. Every plan includes unlimited users, agents and services and all integrations. Approved trials get 1,000 evaluation credits with no time limit.

Limitations to keep in mind

• The lowest paid plan is a significant monthly commitment
• Trial access requires approval rather than instant sign-up
• Write actions must be enabled deliberately, so remediation is not fully automatic out of the box
• Success rates quoted on the site are Cleric's own measurements

Cleric vs alternatives

Few tools in this directory are built for SRE work, so the closest comparisons are coding agents. Devin can debug and open fixes across a repository, which helps once an engineer knows what to look for, but it does not watch alerts. Factory offers a platform of software agents that some teams also point at reliability tasks. Claude Code is useful for an engineer digging through logs and code during an incident. Cleric wins when you want an agent that monitors releases and handles alerts on its own, with predictable per-task costs.

Tips for getting started with Cleric

Connect your observability, deployment and source control tools with read-only access first and let Cleric run on one service. Compare its investigations with your team's own conclusions for a couple of weeks before enabling write access. Route only one alert channel to it at the start so you can measure how many pages it closes. Keep an eye on credit use per investigation to pick the right plan size.

Cleric FAQ

How much does Cleric cost?

Cleric starts at $700 per month on the Team plan, which includes 1,400 credits, as of October 2026. Pro is $1,500 and Scale is $3,000 per month.

Is there a free trial of Cleric?

Yes. After your sign-up is approved, Cleric gives you 1,000 evaluation credits with no time limit.

Is Cleric safe to connect to production?

Cleric is read-only by default, logs every action, requires you to enable write access explicitly, and is SOC 2 Type II compliant.

What is Cleric used for?

Cleric is used to verify releases in production, triage and investigate alerts, and propose fixes so on-call engineers only get paged when a person is needed.