Tokencost Dev
Look up LiteLLM-backed model prices, estimate token costs, and compare models before you commit spend on an agent or SaaS feature.
Overview
io.github.jeff-atriumn/tokencost-dev is a MCP server for the Validate phase that provides LiteLLM-backed LLM model lookup, cost estimation, and comparison for agents.
What is this MCP server?
- LLM pricing oracle with model lookup, cost estimation, and comparison via LiteLLM
- npm package tokencost-dev v0.1.3 with stdio MCP transport
- Supports agent workflows that need grounded per-model token economics
- Open-source repository at atriumn/tokencost-dev on GitHub
- Published npm version 0.1.3 with identifier tokencost-dev
- Three capability areas stated: model lookup, cost estimation, and comparison via LiteLLM
Community signal: 1 GitHub stars.
What problem does it solve?
You cannot trust ad-hoc model cost guesses when choosing LLMs for a solo SaaS or agent, and pricing pages change too often to paste manually.
Who is it for?
Indie builders picking models and estimating inference spend before and during agent feature development.
Skip if: Teams that need enterprise billing reconciliation, GPU cluster costing, or non-LLM infrastructure quotes.
What do I get? / Deliverables
After registration, your agent can quote and compare token costs from a LiteLLM-backed oracle through MCP during planning and implementation.
- Agent-callable model price lookup and comparison
- Token cost estimates grounded in LiteLLM for planning chats
- Repeatable pricing checks across validate, build, and grow decisions
Recommended MCP Servers
Journey fit
Pricing validation happens before full build, but the same oracle supports model choice during agent-tooling and cost review while growing usage. Pricing subphase is the canonical shelf for cost estimation and model comparison gates before prototyping or shipping LLM features.
How it compares
LiteLLM pricing oracle MCP server, not a hosted observability or Prometheus monitoring integration.
Common Questions / FAQ
Who is tokencost-dev for?
Solo and indie builders using coding agents who need accurate LLM price lookup and comparison without maintaining private price tables.
When should I use tokencost-dev?
Use it when scoping LLM features, comparing models for an agent, or revisiting cost assumptions as usage grows.
How do I add tokencost-dev to my agent?
Install the npm package tokencost-dev, add the stdio MCP server to Claude Code, Cursor, or Codex, and invoke cost tools during model selection chats.