llmledger
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
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
} |
| logging | {} |
| prompts | {
"listChanged": true
} |
| resources | {
"subscribe": false,
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| find_modelsB | The highest rated models for a kind of work that meet the limits, each at its cheapest listing, with prices per 1M tokens, Arena rating and prompt size. Unrated models follow, cheapest first. |
| compare_modelsB | Two to six models side by side for a kind of work: rating, prices, cost per task (or plan use), and which to pick, with the break-even share of answer tokens for each against the pick. |
| estimate_costA | What a workload costs on each model at its cheapest listing: per call, per day and per 30 days. With cached or cache write tokens, also the cost without the cache and the saving. |
| price_historyA | Every change to a model's listings since 2023 (prices, retirement dates) with the LiteLLM change behind each, and for listings that have not changed yet, the chance of a price change, a retirement date or removal in the next 90 days and year. With a date, each listing's prices that day. |
| suggest_for_projectB | Reads a prompt, guesses its kind of work, and suggests a model and effort level: the best value (the cheapest whose rating the votes cannot tell apart from the top), the top rated and the cheapest rated. |
| not_yet_pricedB | Models Arena rated recently that LiteLLM's catalogue does not price yet, and how soon the catalogue usually prices a model after Arena first shows it, by maker. |
| estimate_timeA | How long one call and the whole job take, as a usual time and a slow day, learned from your own timed runs of that model called that way (log_run). Says so when there are no runs yet. |
| log_runA | Keep a timed run, or every result of a promptfoo results file, in a file on this computer so estimate_time learns from it. Nothing is sent anywhere. |
| forecast_jobB | The whole job quoted from a promptfoo run of a sample of it. For each model in the sample: the time as a usual time and a slow day, the cost at its cheapest listing next to the one the sample used, the chance of a price change in 90 days, how often it was right (Wilson 95% range) and the break-even against the pick. |
| local_or_cloudA | Pay a cloud model per call, or buy a computer and run an open model at home: the cost a month each way, the month buying pays off, and whether the computer keeps up. The local model's speed comes from your own timed runs (log_run); electricity from EIA or your price; the computer's price is yours. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| pick_a_model | Pick a model for a piece of work, with its cost and effort. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| snapshot | What the plug-in knows today: data dates, how many models and rated tasks, and the sources. |
TDQS
Scored across 10 tools
Several tools overlap in purpose: find_models, compare_models, and suggest_for_project all recommend models; estimate_cost, forecast_job, and local_or_cloud all report costs from different angles; estimate_time and forecast_job both cover timing. Descriptions help clarify boundaries, but an agent could still misselect among them.
All names use snake_case, but conventions are mixed: most are verb_noun (find_models, estimate_cost), while others are noun phrases or adjectives (price_history, not_yet_priced, local_or_cloud). The set is readable but does not follow a predictable verb-first pattern.
With 10 tools, the set is well within the ideal 3-15 range for a specialized LLM pricing and forecasting server. Each tool addresses a distinct facet of the domain, and the count does not feel bloated or thin.
The surface covers model discovery, comparison, cost/time estimation, price history, predictive signals, project suggestions, run logging, job forecasting, and local-vs-cloud analysis. Minor gaps exist, such as direct model lookup by name or a full catalogue enumeration, but core workflows are covered.