Oraclaw
This is a decision math toolkit for agents that need to do actual optimization, not just reason about it. It exposes 17 MCP tools for multi-armed bandits (UCB1, Thompson, LinUCB), Monte Carlo simulation, ARIMA forecasting, linear/MIP solving via HiGHS, graph analysis, and genetic algorithms. Every call returns structured JSON in under 25ms with zero LLM cost. The free tier gives you 25 calls per day with no API key for basic tools like bandit selection and Bayesian updates. Premium tools (CMA-ES, constraint solving, anomaly detection) require an API key. Reach for this when your agent needs to pick the best A/B variant, schedule tasks under constraints, run risk analysis, or solve any problem where guessing costs more than computing the right answer.
- GitHub stars
- 13
- tools
- 17
- auth
- Secret needed
- transport
- STDIO
- env var
- 1
- last updated
- Jun 23, 2026
Tools
17 toolsPublic tool metadata for what this MCP can expose to an agent.
optimize_banditUCB1 / Thompson / Epsilon-Greedy arm selection
UCB1 / Thompson / Epsilon-Greedy arm selection
No parameter schema in public metadata yet.
optimize_contextualContext-aware LinUCB bandit
Context-aware LinUCB bandit
No parameter schema in public metadata yet.
optimize_evolveGenetic algorithm for discrete + multi-objective problems
Genetic algorithm for discrete + multi-objective problems
No parameter schema in public metadata yet.
solve_scheduleEnergy-matched task scheduling
Energy-matched task scheduling
No parameter schema in public metadata yet.
score_convergenceMulti-source probability consensus (Hellinger)
Multi-source probability consensus (Hellinger)
No parameter schema in public metadata yet.
score_calibrationBrier + log score for forecaster accuracy
Brier + log score for forecaster accuracy
No parameter schema in public metadata yet.
predict_bayesianBeta posterior update from weighted evidence
Beta posterior update from weighted evidence
No parameter schema in public metadata yet.
predict_ensembleMulti-model consensus + uncertainty decomposition
Multi-model consensus + uncertainty decomposition
No parameter schema in public metadata yet.
plan_pathfindA* + Yen's k-shortest paths
A* + Yen's k-shortest paths
No parameter schema in public metadata yet.
simulate_montecarloSingle-factor Monte Carlo (6 distributions)
Single-factor Monte Carlo (6 distributions)
No parameter schema in public metadata yet.
simulate_scenarioWhat-if comparison + sensitivity ranking
What-if comparison + sensitivity ranking
No parameter schema in public metadata yet.
optimize_cmaesCMA-ES continuous black-box optimization
CMA-ES continuous black-box optimization
No parameter schema in public metadata yet.
solve_constraintsLP / MIP / QP solver via HiGHS (provably optimal)
LP / MIP / QP solver via HiGHS (provably optimal)
No parameter schema in public metadata yet.
analyze_graphPageRank, Louvain communities, bottleneck detection
PageRank, Louvain communities, bottleneck detection
No parameter schema in public metadata yet.
analyze_riskVaR and CVaR (Expected Shortfall)
VaR and CVaR (Expected Shortfall)
No parameter schema in public metadata yet.
predict_forecastARIMA + Holt-Winters time series forecasting
ARIMA + Holt-Winters time series forecasting
No parameter schema in public metadata yet.
detect_anomalyZ-Score + IQR anomaly detection
Z-Score + IQR anomaly detection
No parameter schema in public metadata yet.
Sponsors
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any languageConfiguration
ORACLAW_API_KEYsecretOptional API key for premium tools (CMA-ES, LP/MIP solver, graph analytics, risk VaR/CVaR, forecasting, anomaly detection). Free tier (11 tools) works without it.
Readme
OraClaw
MCP Optimization Tools for AI Agents -- 17 tools (11 free, no key), sub-25ms. Zero LLM cost.
Your AI agent can't do math. OraClaw gives it deterministic optimization, simulation, forecasting, and risk analysis through the Model Context Protocol. Every tool returns structured JSON, runs in under 25ms, and costs nothing to compute.
Details
- Repository
- github.com/whatsonyourmind/oraclaw
- Package
@oraclaw/mcp-server- Transport
- STDIO
- Auth
- Required
- Registry
- MCP registry, active