Build custom agents on Resolve AI
Compose Resolve AI primitives into your own agents and existing workflows. Use context, investigation, and remediation without rebuilding them
Build your own agents on Resolve
Resolve is exposed as MCP, API, and Skills. Your agents call Resolve for production context, investigation, and remediation, without rebuilding the primitives
Root cause: Connection pool exhaustion, not CPU. RDS CPU is ~10% (baseline). All 80 connections on the primary are held by one long-running reporting query (running 8m), blocking new checkout transactions since 14:18 UTC.
- Terminate blocking query pid 18472 to free connections immediately.
- Suggested fix: Add statement_timeout = '30s' to the reporting-service DB role at db/roles/reporting.tf:42.
- Follow-up: rewrite the aggregation in report_engine.go:142 to avoid full scans.
Use Resolve in the workflows your team already lives in
Call Resolve from Slack, your CLI, your IDE, or another agent. Production context and remediation surface wherever your engineers already work
12 customer tickets in 15m: “can’t complete checkout”
support-monitorservice:checkout-servicecustomer-symptom-investigation- ticket_count
- 12
- common_symptom
- “can’t complete checkout”
- sample_ids
- [INC-4821, INC-4824, INC-4829, +2]
- first_seen_at
- 4m ago
Plug your skills into Resolve
Resolve consumes the runbooks, internal tools, and tribal knowledge your team has already built. Every investigation uses what you already have
pg-replication-lag
kafka-consumer-lag
deploy-health-check
rds-cpu-investigation
oncall-handoff
incident-postmortem
Used and loved by engineers
Removing the toil of investigations, war rooms, and on-call
“We pull fewer engineers into war rooms, on-call is materially better, and that translates directly to advertiser trust and revenue protection.”
“I don't need more numbers or more data. What I need is a root cause.”
“Resolve AI proved it could deliver real results in a constrained environment. It identified dependencies, surfaced accurate root causes 73% faster than our teams, all while integrating cleanly into our existing stack.”
“Resolve AI makes our junior on-call engineers as effective as our seniors, flattening the experience curve. We've seen a 2x productivity lift while eliminating the runbook gap.”
Recent updates.
- May 2026
MCP server beta
Call Resolve AI from any MCP-compatible client
- April 2026
Public API
Programmatic access to Resolve AI primitives
Frequently asked questions
Resolve exposes three composable primitives: MCP, API, and Skills. Start with whichever fits your stack. Most teams begin with MCP because their agent framework already speaks it
Runbooks, internal tools, scripts, and knowledge bases. Resolve uses what your team has already built rather than asking you to rebuild it
Through your existing auth and permissions. Resolve does not replicate your stack, it composes with it
The custom agent layer is for engineers. The agents you build with it work for everyone you support
The same as the rest of Resolve: read-only by default, scoped action permissions, and an audit log on every call
Talk to us. We support several deployment models depending on your environment