Blog
Notes from the loop
Product updates, engineering deep dives, and research from the team building cyql.
Stop repeating the same mistake
Failure clustering is now generally available. Recurring agent failures surface as a pattern — and one line in your config makes them stop.
Introducing cyql: the agent platform that ships code
Why we built an autonomous platform that owns a task end to end — from plain-language request to a reviewed pull request.
What teams across fintech, dev tools, and infra are actually shipping
A look at the real work cyql handles every day — and the patterns that keep showing up across industries.
The agent fleet that scales itself
Why we use queue depth as the scaling signal — and what it means when a large refactor finishes in minutes instead of hours.
How we isolate autonomous agents
Pod-per-task isolation, locked-down egress, and short-lived credentials — the security model that lets you say yes to autonomy.
From issue label to merged PR
cyql now picks up Linear and Jira issues automatically and closes the loop back to the ticket. Your issue tracker becomes the interface.
Agents that review agents
A single agent is a draft. A review loop is a product. How cyql turns raw generations into PRs that pass human review.
Zero standing secrets
Long-lived credentials and autonomous agents are a dangerous combination. Here is how we removed them entirely.
Every prompt, every diff, one timeline
Execution trace viewer 2.0 puts the full run on a single scrollable timeline — with deep links to any step.
Execution traces and failure clustering: closing the feedback loop
Why observability is not optional for autonomous agents — and how replay and clustering turn failures into guardrails.
Building the loop: durability, isolation, and why both matter
How we designed the plan→ship loop and why making it durable and isolated was not optional.
Why we started cyql
Software teams don't have a code-writing problem. They have a follow-through problem. We built cyql to close the gap.