Zapvol Documentation

Set up Zapvol and run your first agent task in minutes.

Design principles and mental models — how to think about Zapvol's primitives before reading the architecture.

How the monorepo, packages, and runtime pieces fit together.

Cross-cutting platform concerns shared by every subsystem — authentication, storage, secrets, networking.

Observability stack, log pipeline, dashboards — how to run and monitor Zapvol in production.

The economics of LLM products. Cost model, control mechanisms, and ROI framework — universal thinking for any LLM agent product.

Planned — business metrics for LLM agent products: north-star, retention, unit economics, activation path.

Planned — pricing strategy for LLM agent products: subscription, usage-based, hybrid, tier design, enterprise contracts.

Planned — growth mechanisms for agent products: activation, expansion, self-serve vs sales-led, referral, onboarding.

Planned — operational playbooks for customer success, onboarding, expansion, churn-save, and internal support.

Planned — data compliance, audit trails, model output liability, and industry certifications (SOC2 / HIPAA / GDPR).

Vercel AI SDK internals as they matter for building production agents — three-layer API, lifecycle ordering, message reference model, prepareStep semantics. Based on ai@^6.

Build agent harnesses — long-running flows, multi-agent coordination, context engineering.

Design theory for prompts, memory, and context lifecycle — treat the attention budget as a finite resource.

Reverse-engineer Claude Code — the reference agent harness — to extract design principles for building your own.

Author and ship reusable skills that extend the agent's capabilities.

The Browser Use Agent — remote-control the user's logged-in browser under per-domain consent.