HTMLBrowser-local developer utilities
Measure the hidden weight in your AI stack.
Inspect MCP schemas, move configs between clients, expose tool collisions, build rollout plans, and budget an agent's context window. Your inputs stay in this browser tab.
MCP payload analysis
Tool-schema budget analyzer
Paste a tools/list response or raw tool array. See what every tool costs before a model reads a single result.
Run an analysis to rank expensive schemas and surface catalog quality issues.
Portable MCP setup
MCP config converter
Normalize common mcpServers JSON and generate generic JSON or Codex TOML without copying credentials into the result.
Unsafe output enabled. Exported text may contain live credentials. Review it before copying or downloading.
Catalog conflict analysis
Tool collision detector
Compare named catalogs for exact names, confusing near-matches, and descriptions that promise substantially the same action.
Evidence for each collision will appear here with similarity scores.
Activation readiness
AI stack readiness audit
Paste an MCP config and a tool catalog to get a launch scorecard, fix plan, evidence checklist, and copyable handoff brief.
Build a launch plan to produce readiness, risks, verification steps, and evidence to save.
Context-window planning
Agent context budget calculator
Reserve the answer first, then see how much room remains for tools, instructions, system policy, and the active conversation.
How estimates work
Useful signal, without pretending tokenizers are identical.
Schema estimates use UTF-8 bytes divided by four—a practical planning approximation, not a provider bill. Collision scores combine normalized names and description-word overlap. Verify exact production budgets with the tokenizer used by your target model.