How it works
Understand the local proxy pattern, request flow, provider routing, caching behavior, and where TTL fits.
← Back to home
Resource library
Explore the proof, setup, and support pages.
Use this page as the organized map for AI Optimizer. Start with cache proof, move into setup, then go deeper on provider support, developer workflows, and trust questions.
Quick answer
AI Optimizer is easiest to evaluate in this order: understand the repeated-request claim, inspect the cache proof, review the local setup path, then check whether your provider and workflow fit the product's strongest lane.
Start here
These pages explain the product, the setup path, and the trust questions most visitors should answer first.
Install guide
Set up the desktop app, start the local proxy, and route compatible traffic through http://localhost:3000/v1.
FAQ
Review support scope, provider fit, privacy, caching behavior, and the most common adoption questions.
Proof and cache behavior
The strongest pages for judging whether the repeated-request caching claim is credible.
Cache proof
See the practical proof path for repeated requests, cache hits, TTL behavior, and provider support.
Exact cache-hit proof
Go deeper on exact local cache hits and why visible repeat behavior matters for trust.
Exact hits vs partial hits
Understand the difference between local exact-hit caching and provider-side prompt reuse signals.
Reused prompt tokens
See how provider-side token reuse differs from the local exact-hit lane AI Optimizer makes visible.
Cache OpenAI locally
Review the OpenAI-focused local caching path for repeat-heavy scripts and workflows.
Install and proof video
Watch the install path and a repeated request proof flow in one short walkthrough.
Provider cost pages
Provider-specific pages for OpenAI, Anthropic, and Google Gemini support.
Reduce OpenAI costs locally
Use a local-first cache and control layer for repeat-heavy OpenAI workflows.
OpenAI CLI workflows
Route compatible command-line workflows through localhost while keeping the terminal workflow familiar.
Reduce Anthropic costs
Review the focused Anthropic chat-completions lane for repeated Claude workflows.
Reduce Gemini costs
Understand the narrower Gemini generateContent support lane and where it fits best.
Developer and workflow guides
Use these when you already understand the wedge and want practical workflow examples.
For developers
Developer-focused setup patterns, base URL changes, and repeat-heavy local workflow examples.
For agents
Agent workflow examples where retries, loops, and repeated tool calls can create quiet duplicate spend.
Scripts, cron, and agents
See why scheduled prompts, scripts, and recurring automations are a clean fit for local caching.
Monitor local API usage
Track local request behavior, cache-hit patterns, and repeat-heavy OpenAI workflows.
Add a local proxy
Use the common localhost base URL pattern without rebuilding the whole workflow.
Prompt testing
Reduce waste during repeated prompt testing, reruns, and iteration-heavy local workflows.
Trust and company pages
Use these pages to verify product scope, privacy posture, and the local-first philosophy behind AI Optimizer.
About AI Optimizer
Read the product philosophy: local-first, proof-led, and strongest where repeated AI work is real.
Privacy policy
Review how the site and product present privacy, personal information, and data handling.
FAQ
Check the practical trust questions before trying the product in a real workflow.
Start with proof, then choose your path.
If the repeated-request cache proof makes sense, install AI Optimizer and test it on one workflow before expanding further.