Resources
Free references for teams shipping AI
Practical tools and references — the kind of thing you reach for when you’re lost mid-task and there’s no in-house expert to ask. When you need someone to ship it, that’s what we do.
Guides
16 guidesPractical, engineer-written answers to the questions teams actually ask — reducing LLM costs, the Model Context Protocol, building evals, and agentic coding.
OpenVendor comparisons
6 listsBuyer-fit comparisons of AI engineering firms and products — what they do, who they suit, what they cost. Not rankings; our own fit disclosed separately.
OpenPlaybooks
8 playbooksLong-form, senior-engineer-grade how-to for the AI engineering work we do.
OpenReference architectures
3 reference architecturesWhat a typical engagement looks like end-to-end — week-by-week plan, deliverables, end state, ownership.
OpenProductionize AI code
4 tool playbooksTake an AI-generated prototype (Lovable, Bolt, v0, Replit Agent) to production — per-tool playbooks and the 12-axis hardening rubric.
OpenLLM cost calculator
Interactive toolEstimate and compare your monthly token bill across every model.
OpenLLM pricing comparison
19 modelsCost per million tokens and context windows, side by side.
OpenMCP integration guides
25 integrationsConnect Claude Code and Cursor to your internal tools, safely.
OpenTool comparisons
12 comparisonsHonest "X vs Y" breakdowns of the tools you’re evaluating.
OpenAI engineering glossary
40 termsPlain-English definitions of the terms behind production AI.
OpenLLM error fixes
16 fixesStep-by-step fixes for the errors that block AI work.
OpenWhich of these you want, by what you are trying to do
There are eight clusters here and they answer different questions, so the fastest route in is by intent rather than by browsing. If you are trying to understand something well enough to make a decision, read a guide. If you have decided and need a sequenced plan with weeks attached, read a playbook. If something is broken right now, go to the error fixes, which are written cause-first rather than as tutorials. If you need a term defined precisely for a document or a conversation, the glossary is faster than any long-form page.
If you are choosing between tools, the comparison pages take a position rather than listing features in two columns, and each one names where the option we did not pick genuinely wins. If you are choosing a supplier, the best-of lists are honest about the fact that we appear in some of the categories. If you are budgeting, the cost calculator gives a shaped estimate and the pricing comparison carries a verification date, because model prices move faster than any page should pretend to track.
Everything here is free, un-gated, and written to be quoted. There is no email wall, no download form, and no chapter withheld to force a conversation. The commercial logic is straightforward and we would rather say it than dress it up: teams who find these useful sometimes hire us, and teams who use them to do the work themselves are a perfectly good outcome. Several of the playbooks describe engagements we sell, in enough detail for a competent team to run them without us.
Nothing here is generated and published unreviewed. Pages carry the date they were last checked, ranges drawn from our own engagements are labelled as such, and where we cannot source a claim about another company's pricing or licensing we describe the shape of the offering instead of inventing a figure. If something has gone stale or is simply wrong, telling us costs you a minute and improves the page for everyone who finds it next.