Everything you need to track and optimize AI costs
Guides, documentation, glossaries, and tools built for engineering teams who want full visibility into their AI spend without the guesswork.
Explore
Guides & References
Whether you're getting started or optimizing at scale, these resources cover every aspect of AI cost management.
AI Cost Glossary
The definitive glossary of AI cost management terms. Tokens, billing, optimization, model routing, and more.
ExploreGetting Started
Set up CostHawk in under five minutes. Connect your first provider and start tracking AI spend immediately.
Open docsIntegration Guides
Step-by-step guides for connecting OpenAI, Anthropic, Google, and other providers via admin sync, MCP, or wrapped keys.
Open docsModel Pricing
Search and compare indexed pricing for models across 12+ provider families. Updated automatically from CostHawk pricing tables.
ExploreMCP Server
Local-first telemetry for Claude Code, OpenAI Codex CLI, and OpenCode. Track usage without provider admin API access.
Open docsProvider Comparisons
Side-by-side comparisons of pricing, rate limits, and capabilities across major AI providers.
Cost Optimization Guides
Practical playbooks for reducing AI spend through prompt engineering, model routing, caching, and context management.
Glossary Spotlight
Five concepts that explain most AI cost surprises
These are the terms teams usually need right after pricing tables: routing, telemetry, output ceilings, runtime architecture, and the metadata you keep.
OpenTelemetry
Understand OTLP traces, metrics, and logs well enough to wire real-time AI telemetry without losing cost visibility.
Read the glossary entryMax Tokens
Use output ceilings to control runaway completions, latency, and cost per request before those limits show up in invoices.
Read the glossary entryServerless Inference
Compare pay-per-request inference economics against self-hosted GPU stacks when deciding how to run production traffic.
Read the glossary entryLogging
Track the request metadata that actually explains token spikes, latency regressions, and expensive failure modes.
Read the glossary entryLLM Gateway
See how centralized routing layers help with policy enforcement, provider failover, attribution, and spend control.
Read the glossary entrySupported providers
12+
Provider families with indexed pricing
Tracking paths
3
MCP telemetry, admin sync, wrapped keys
Setup time
< 5 min
Connect your first provider
Resources
Ready to take control of your AI costs?
Start with local telemetry or plug in the tracking path your team needs. CostHawk gives you one view of usage, cost, and attribution across your entire AI stack.