Olypsis / Resources
A useful place
to start.
Work through a question. See what we learned.
Try the numbers for yourself.
Tools
Try your own numbers. See what to check next.
Practical guides
Pick the task you need help with. Each guide gives you steps to try.
Why does my AI agent cost more on each turn?
Trace repeated context, large tool results, cache misses and retries before choosing a cost fix.
Should I trim context or switch to a cheaper model?
Choose a cost experiment from the evidence: repeated input, large tool results, expensive output or extra retries.
How do I test a cheaper AI agent without losing quality?
Build a repeatable comparison of task quality, complete-run cost, latency and review effort before a rollout.
How to evaluate an AI project before you build
Assess the workflow, data, expected benefit, and risks of an AI project. Define the evidence you need before committing to a build.
How we scope and price an AI project
What goes into a proposal, what affects the price, and how to plan for costs after launch.
How to plan human review for AI agents
How to divide a task into useful steps, test the results, and decide where approval is needed.
From the blog
What we learn as we build. Read the ideas behind the work.
Visit the blog ↗Stop wasting tokens. Start with these four checks.
A per-turn checklist for AI agents: inspect the system prompt, trim tool output, verify caching, and choose a model that fits the job.
Give your first AI agent a smaller job.
How to choose a useful starting point, define a finished result, and make the first deployment worth expanding.
What an AI agent needs before it goes live.
The engineering decisions that let a team trust an agent with everyday work, including the days something goes wrong.
The real cost of running an AI agent.
Model usage is one line in the budget. A useful cost model follows the whole task through review, retries, and completion.
Case studies
The work, the result and the limits of what it tells us.
Explore all client work ↗