ModelCostSaver
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MODELCOSTSAVER_LEDGER | No | Enable the local record_usage write. Default is 'off'. | |
| MODELCOSTSAVER_REFRESH | No | Enable the opt-in remote catalog refresh. Default is 'off'. | |
| MODELCOSTSAVER_PROVIDER | No | Bias select_optimal_model toward a specific provider. Default is none. | |
| MODELCOSTSAVER_PROVIDERS | No | Allowlist for recommendations (e.g., 'anthropic,openai'). Default is client-derived. | |
| MODELCOSTSAVER_TELEMETRY | No | Telemetry setting (kept off for transparency). Default is 'off'. | |
| MODELCOSTSAVER_FAST_MODEL | No | Pin a preferred model for the fast tier. Default is catalog cheapest. | |
| MODELCOSTSAVER_CATALOG_URL | No | Override the refresh source URL. Default is bundled. | |
| MODELCOSTSAVER_INCLUDE_LOCAL | No | Surface self-hosted / $0 models. Default is 'off'. | |
| MODELCOSTSAVER_TRIVIAL_MODEL | No | Pin a preferred model for the trivial tier. Default is catalog cheapest. | |
| MODELCOSTSAVER_STANDARD_MODEL | No | Pin a preferred model for the standard tier. Default is catalog cheapest. | |
| MODELCOSTSAVER_CHARS_PER_TOKEN | No | Tune the token estimator. Default is '4'. | |
| MODELCOSTSAVER_REASONING_MODEL | No | Pin a preferred model for the reasoning tier. Default is catalog cheapest. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| prompts | {
"listChanged": true
} |
| resources | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| estimate_costA | Estimate the cost of a single LLM call for one model from known or estimated token counts. Offline, no keys. |
| predict_costA | Forecast the cost of a prompt across candidate models before the call. Returns a cheapest-first ranking with assumptions. Offline. |
| select_optimal_modelB | Pick the single cheapest model that meets the task tier, capabilities, and budget, with full reasoning and a fallbackChain. Offline. |
| compare_modelsB | Compare models side by side for a fixed token shape, cheapest first, with the multiple of the cheapest. Offline. |
| list_modelsB | Return the model catalog with pricing, optionally filtered. capabilities are arrays. Offline. |
| get_pricingC | Return the model catalog with pricing, optionally filtered. capabilities are arrays. Offline. |
| optimize_requestA | Check whether a cheaper capable model exists for a call you plan to make, and report the savings. Offline. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| modelcostsaver | How and when to use ModelCostSaver: forecast cost and pick the cheapest capable model before an LLM call. |
| modelcostsaver-setup | Self-configure ModelCostSaver in this client. Hand this to your agent to set it up. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| catalog | The full model catalog with per-token pricing, tiers, and capabilities. Offline, versioned. |
TDQS
Scored across 7 tools
Several tools have overlapping purposes (e.g., compare_models, predict_cost, optimize_request, select_optimal_model all compare or select models based on cost). Additionally, get_pricing and list_models have identical descriptions, causing ambiguity.
All tool names follow a consistent verb_noun pattern using snake_case (e.g., estimate_cost, list_models, select_optimal_model).
7 tools is well-scoped for a model cost estimation server, covering the essential operations without being excessive.
The tools cover listing, pricing, cost estimation, comparison, and optimization. However, the presence of a duplicate tool (get_pricing and list_models) indicates a minor gap in tool design.