Reviewmetric
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| REVIEWMETRIC_API_KEY | Yes | Your API key from the web dashboard |
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": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| help_meA | Show everything ReviewMetric can do in Claude, what costs an RM token, and example requests. |
| submit_manuscriptA | Submit a local .docx manuscript to the ReviewMetric review panel (uses 1 RM token). A panel of specialist AI referees checks novelty against the published literature, whether the results support the claims, methods and reproducibility, citation accuracy (including retracted references), figures, and clarity, then a handling editor writes prioritised revisions and a referee-style report. A review usually takes about 5 minutes. Args: file_path: Absolute path to the .docx file on this computer. |
| check_analysis_statusA | Check a ReviewMetric review. Returns progress while running; when done, the full report, the manuscript's place on the ReviewMetric Atlas, and how many text fixes are available. Args: run_id: The Run ID returned by submit_manuscript (or listed by list_my_reviews). |
| get_report_jsonB | Return the complete ReviewMetric report as JSON (all findings, literature, reference checks). Args: run_id: The Run ID of a completed review. |
| list_my_reviewsA | List the user's ReviewMetric reviews, newest first, with their scores and Run IDs. Args: limit: How many reviews to list (1 to 50). |
| get_text_fixesA | Text fixes of a completed review: every change that needs no new experiment (wording, overstated claims, citations, missing references, method details, typos). Free preview before they are prepared; once prepared, the exact edits (where, what to find, what to write) that you can apply to the user's manuscript. Args: run_id: The Run ID of a completed review. |
| prepare_text_fixesA | Prepare the text fixes of a completed review (uses 1 RM token; asking again for the same review is free). Only call this when the user asked for the fixes. Then call get_text_fixes to receive the edits. Args: run_id: The Run ID of a completed review. |
| search_atlasA | Search the ReviewMetric Atlas, the map of landmark discoveries since 1900, their Nobel Prizes and today's frontiers. Matches names, topics and summaries. Args: query: Words to look for, e.g. "CRISPR", "superconductivity", "neural networks". field: Optional field: Physics, Materials, Chemistry, Biology, Medicine or Computer science. limit: How many results (1 to 20). |
| list_frontiersB | The frontiers of the ReviewMetric Atlas: open problems at the edge of science, what each one builds on, and the research groups pushing it now. |
| place_on_atlasA | Where one of the user's reviewed manuscripts sits on the ReviewMetric Atlas: its topic, the landmarks it cites, the closest discoveries, and the frontier ahead of it. Args: run_id: The Run ID of a completed review (see list_my_reviews). |
| atlas_whats_newB | This week's additions to the ReviewMetric Atlas (updated every week). |
| trace_pathB | The chain of discoveries linking two stars of the ReviewMetric Atlas, e.g. from the double helix to AlphaFold, or from the Turing machine to the fault-tolerant quantum computer. Args: from_discovery: Name (or Atlas id) of the first discovery. to_discovery: Name (or Atlas id) of the second discovery or frontier. |
| unlock_atlas_sourcesA | Unlock the Atlas sources for 30 days: the leading paper and research group behind every discovery, and who is pushing each frontier. Costs 1 RM token (added to any current access). IMPORTANT: first call without confirm to get the cost and the user's balance, tell the user, and ask them. Call with confirm=true ONLY if the user then explicitly answered yes in this conversation. Args: confirm: True only after the user's explicit yes. |
| account_statusA | The user's ReviewMetric account: RM tokens left, Atlas access, number of reviews, and prices. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| help_me | What can ReviewMetric do for me? |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 14 tools
Most tools target clearly distinct resources/actions (submit_manuscript, search_atlas, place_on_atlas, trace_path, prepare_text_fixes). The main overlap is between check_analysis_status and get_report_json, both keyed on run_id and both returning report content, which could cause misselection when a review is complete.
The dominant pattern is consistent snake_case verb_noun (submit_manuscript, list_my_reviews, get_text_fixes, search_atlas). A few deviations exist (help_me, account_status, atlas_whats_new with an unusual noun-first ordering), but they are minor and still readable.
14 tools is well within a coherent range for a review-plus-atlas service. Each tool maps to a distinct workflow step (submit, monitor, report, fixes, atlas exploration, account) and none feel redundant padding.
The surface covers the full lifecycle: submission, status polling, structured report retrieval, text-fix preview/prepare, atlas search/navigation, and account/help. Minor gaps exist (e.g. no cancel/delete review or export/share operation), but core workflows are fully covered.