0xwilliamortiz
Featured
openclaude-improved
OpenClaude is an open-source CLI coding agent that runs on any platform and supports a wide range of LLM providers, offering the same tools and workflows as Claude Code.
bindwidth
A frontend-only, no-telemetry calculator that sizes a private local AI stack by finding which of three constraints—KV cache memory, combined prefill/decode serving capacity, and runtime session ceiling—actually binds.
Stars
55
7-day growth
No data
Forks
1
Open issues
0
License
MIT
Last updated
2026-08-01
AI repository intelligence
FR-AI / ANALYSISWhy it is worth attention
It directly addresses common sizing mistakes by showing the binding constraint, comparing owned hardware, sovereign rental, and enterprise API options, and separating measured data from estimates so organizations can make an explicit, reviewable decision record.
Who it is for
- Teams evaluating whether to run LLMs locally for data governance
- Infrastructure engineers sizing GPU or accelerator hardware
- Technical and financial planners comparing build vs rent vs subscribe
- AI agent platform operators needing capacity estimates
Use cases
- Estimate hardware needed to serve both interactive users and autonomous agents with a local LLM
- Compare owned hardware, sovereign rental, and enterprise API subscriptions across cost and capability
- Generate a Markdown or JSON decision record for procurement review
- Identify which constraint—KV memory, serving capacity, or session ceiling—binds before buying
Strengths
- Runs entirely in the browser with no server, database, account, or telemetry; typed data is not transmitted
- Explicitly shows which constraint binds for every configuration, preventing purchases against the wrong limit
- Separates measured from estimated data and includes confidence levels in exported reports
- Exports a Markdown decision record and a reloadable JSON scenario for transparency
Considerations
- It is a directional sensitivity estimator, not a procurement quote; exact model, runtime, and hardware must be validated before purchase
- It does not tell you whether a premium is worth paying or establish that subscription and local models are equivalent in quality, latency, tools, or availability
- It requires serving over HTTP because opening index.html directly from the filesystem blocks fetch on file://
README quick start
Quick start
Description
Evidence-aware on-prem LLM inference sizing and TCO calculator
Related repositories
Similar projects matched by category, topics, and programming language.
deerwork-ai
Featured
deer-workflow
An open-source Dynamic Workflow runtime that combines deterministic TypeScript orchestration with replaceable Agent runtimes.
0xwilliamortiz
Featured
ratchet
Ratchet installs coding-agent hooks that measure every edit against a minimalism ruleset, report findings back mid-session, and enforce budgets through advise, guard, or strict modes.