About ExcelliaAI MCP
ExcelliaAI MCP is an MCP server in the Documentation category: aI-Powered Spreadsheet Validation & Data Quality Platform Excellia AI is a local-first, AI-driven spreadsheet validation platform built to automate data cleaning, validation, anomaly detection, and enrichment for large datasets. It combines rule-based checks, machine learning, and local LLM agents to drastically reduce manual analyst effort.
Install
git clone https://github.com/AnuragRoque/ExcelliaAI-MCPThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Tools (17)
ask_dataChat with your data, hallucination-proof: the LLM plans a query, pandas computes it, and the evidence table always comes back with the answer
Parameters not published.
dedupe_rowsEntity resolution: cluster near-duplicates, keep a canonical row, log every merge
Parameters not published.
detect_anomaliesIsolation Forest + column outliers + rare categories + near-duplicates + pattern breaks — every flag has a confidence and a reason
Parameters not published.
evaluate_fraud_modelLabelled holdout → real-world accuracy next to the training metrics
Parameters not published.
export_reportHighlighted xlsx report + Data Health Score with its full deduction breakdown
Parameters not published.
job_statusPoll long-running work started with `async_=true` (big files run on a job queue)
Parameters not published.
list_fraud_modelsEvery saved ModelCard (metrics and features — never your data)
Parameters not published.
match_namesKYC hybrid name matching — deterministic similarity, optional offline-LLM verdicts
Parameters not published.
profile_sheetRow/column counts, inferred types, null rates, stats, auto-detected formats
Parameters not published.
reconcileMatch two spreadsheets by key columns; tolerances for amounts, date windows, fuzzy names; four buckets incl. field-level discrepancies
Parameters not published.
Show all 17 tools
run_recipeReplay a saved cleanup on next month's file in one call
Parameters not published.
save_rulesetPersist a custom validation ruleset (also readable as the `ruleset://` MCP resource)
Parameters not published.
score_fraudFresh file → per-row fraud probability, risk band, and the top factors pushing each score up
Parameters not published.
train_fraud_modelLabelled history → risk model with an honest ModelCard (CV metrics, leakage detection, refusals that name the fix)
Parameters not published.
transform_applyApply a recipe to a NEW file (originals are never touched); new values land in `_ai`-suffixed columns unless you say replace
Parameters not published.
transform_previewInstruction → cleaning recipe + before/after on a 20-row sample. Nothing changes yet
Parameters not published.
validateDeterministic rule checks: required fields, GST/PAN/Aadhaar/email/phone/IFSC formats, ranges, duplicates, mixed types. Rulesets: `default`, `kyc`, `invoice`, `payroll`, `bank-statement`, plus your own saved ones
Parameters not published.
Ask AI
Ask AI about ExcelliaAI MCP
Powered by Claude · Grounded in docs
Security checks
6 checks have not been run yet.