AI Impact MCP
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
} |
| resources | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| estimate_impactA | Estimate the environmental impact (energy kWh, miles driven in a gas car, water for cooling, CO2e) for a single AI request given its token counts. Uses the EcoLogits life-cycle methodology. |
| log_usageB | Record one AI request's token usage into the local store so it shows up in reports. Use this to manually log usage from any client. |
| reportC | Summarize recorded AI usage and its environmental impact over a period (today, week, month, all), broken down by model. |
| efficiency_scoreA | Score how efficiently a conversation was set up (fewest prompts/rework). Pass the conversation turns. Returns a 0–100 score, grade, and actionable tips. |
| analyze_efficiencyA | Run the efficiency coach over your most recent Claude Code sessions (reads transcript text on-demand, never stores it). Returns per-session scores, an average, wasted-rework tokens, and your top recurring tips. |
| set_scenarioA | Set the default confidence scenario for future estimates. conservative = lowest (min active params), midpoint = mean, high = max. |
| scan_logsA | Backfill exact AI usage from Claude Code's local session transcripts (~/.claude/projects). Reads only token counts + metadata, never message content. Idempotent — safe to run repeatedly. |
| record_web_chatA | Record ESTIMATED usage for a Claude desktop/web conversation that doesn't expose token counts. Preferred: pass structured |
| generate_dashboardB | Build a standalone HTML dashboard (charts of energy/carbon/water over time and by model) from your recorded usage. Returns the file path to open in a browser. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| methodology |
TDQS
Scored across 9 tools
Most tools have clearly distinct purposes (e.g., estimate_impact vs. log_usage), but some overlap exists: efficiency_score and analyze_efficiency both address efficiency, and record_web_chat and log_usage both handle logging. Descriptions mitigate confusion.
Names consistently use lowercase with underscores (e.g., analyze_efficiency, set_scenario). Most follow a verb_noun pattern, but efficiency_score is noun_verb, and verbs vary (analyze, estimate, log, scan). Minor inconsistency.
9 tools is well-scoped for tracking AI usage, estimating environmental impact, and reporting. Each tool has a clear role, with no obvious bloat or deficiency.
Covers core logging (manual, web, session scans), estimation, efficiency analysis, reporting, and scenarios. Missing CRUD for logged records (e.g., delete or edit) and export, but primary workflows are present.