Blog
Code is nearly free now. Infrastructure is the product.
Six posts for the launch: the announcement, the two comparisons, and one for each door. Earlier posts and playbooks come back rewritten.
- Announcement · Post 1 of 6
Introducing Paddington: sovereign agentic infrastructure
Code is nearly free now. What a company owns is the infrastructure its agents run on, and the people who design it, build it, trust it and run it. That is what we build.
5 min read - Comparison · Post 2 of 6
Why Kilter is easier than Lovable
Lovable makes the first screen easy. Everything after the first screen is infrastructure, and that is where Kilter does the work for you.
6 min read - Comparison · Post 3 of 6
Better than Supabase: agentic infrastructure is vastly more than Postgres
Supabase is a very good Postgres with services around it. Agents need a whole platform: workflows, queues, memory, a model gateway, policy, identity for non-humans, and somewhere to run for hours.
6 min read - Method · Post 4 of 6
Going fully agentic, one value stream at a time
Company-wide AI programmes stall because they start from tools. Start from the work: one stream from request to outcome, agents on it with the people who run it, then the next.
6 min read - Platform · Post 5 of 6
Stop piloting. Start performing.
A pilot proves a model can do a task. Production proves the company can run it. The gap between them is platform work, and it is the same gap for every pilot.
6 min read - Business · Post 6 of 6
Keep the models. Own the workspace.
Claude for Teams and Enterprise put agents on every desk. The workspace around them, where the artifacts live, which tools agents can call and what gets recorded, should be yours. Pucks, KilterDesktop and managed MCP on rails make it yours.
7 min read
Newsletter
New posts, by email.
A short note when we publish. No tracking pixels, and you can leave any time.