Free tools for building better AI agents.
Design, validate, evaluate, calculate, and operate AI-agent systems with practical browser-based utilities. No API key required.
Featured tools
High-value utilities for designing and hardening real agent systems.
AI Agent Architecture Builder
Turn use-case requirements into a practical agent architecture, component map, and build checklist.
Open tool →Agent Loop Detector
Inspect an agent trace for repeated calls, oscillating patterns, repeated failures, and runaway-step risk.
Open tool →RAG Benchmark Calculator
Calculate Precision@K, Recall@K, Hit Rate, MRR, and nDCG from relevant IDs and a ranked retrieval result.
Open tool →Tool Contract Linter
Lint tool contracts for vague names, weak descriptions, schema mistakes, overlapping tools, and side-effect ambiguity.
Open tool →Agent Memory Architecture Builder
Choose working state, session memory, durable profile memory, episodic history, semantic knowledge, and retention policies.
Open tool →Agent Production Readiness Scorecard
Score architecture, state, tools, safety, evaluation, observability, reliability, and operating readiness.
Open tool →All AIRundown tools
Search by problem, or filter by the part of the agent lifecycle you are working on.
AI Agent Architecture Builder
Turn use-case requirements into a practical agent architecture, component map, and build checklist.
Open tool →AI Agent Stack Recommender
Choose a sensible implementation stack for agents, RAG, memory, tools, observability, and deployment.
Open tool →Agent vs Workflow Decision Tool
Decide whether you need automation, an LLM workflow, a single agent, or multiple agents.
Open tool →LLM Cost Calculator
Estimate per-request, daily, and monthly model cost using the pricing you enter.
Open tool →RAG Architecture Builder
Design ingestion, indexing, retrieval, reranking, access control, and evaluation around your data.
Open tool →Function Schema Builder
Build a clean JSON tool/function schema from a simple parameter definition.
Open tool →Tool Schema Compatibility Checker
Inspect one tool schema and flag structures that may need review across MCP and common function-calling formats.
Open tool →MCP Tool Schema Footprint Analyzer
Measure tool count, JSON size, estimated context footprint, duplication, and oversized schemas.
Open tool →RAG Chunk Visualizer
Paste real text and visualize approximate token-based chunks and overlap before you index.
Open tool →RAG Context Budget Calculator
Allocate context among system instructions, history, tools, retrieval, output reserve, and other payloads.
Open tool →Agent Trajectory Visualizer
Turn a list of agent steps or tool calls into a readable execution path.
Open tool →Trajectory Diff Evaluator
Compare expected and actual agent trajectories using strict, ordered-subset, or unordered matching.
Open tool →Agent State Machine Designer
Define states and transitions, visualize the flow, and export Mermaid and JSON.
Open tool →Tool Permission Matrix Builder
Map read, write, delete, external side effects, and human approval across an agent toolset.
Open tool →Agent Permission Risk Calculator
Estimate agency risk from powerful permissions and identify where guardrails or approval are needed.
Open tool →Agent Loop Cost & Latency Simulator
Estimate task cost and latency across iterations, model calls, tool calls, retries, and parallel execution.
Open tool →Tool Call Debugger
Compare actual tool arguments against a JSON schema and flag missing, unknown, or mistyped fields.
Open tool →Vector Storage Calculator
Estimate raw vector storage plus metadata and index overhead from count, dimension, and precision.
Open tool →Retry & Backoff Calculator
Preview fixed, linear, or exponential retry schedules with caps and optional jitter ranges.
Open tool →Agent Production Readiness Scorecard
Score architecture, state, tools, safety, evaluation, observability, reliability, and operating readiness.
Open tool →Agent Timeout Budget Planner
Allocate an end-to-end task timeout across model turns, tools, retrieval, approvals, and retry headroom.
Open tool →Agent Error Recovery Policy Builder
Turn common failure classes into deterministic retry, fallback, escalate, or stop rules.
Open tool →Idempotency & Side-Effect Safety Checker
Review duplicate-action risk for writes, payments, messages, deletes, and other side-effecting tools.
Open tool →MCP Capability Planner
Choose the MCP primitives a client/server integration actually needs: tools, resources, prompts, sampling, roots, elicitation, logging, progress, and cancellation.
Open tool →MCP Production Readiness Checker
Score an MCP integration across contracts, authorization, consent, timeouts, logging, errors, approvals, and least privilege.
Open tool →Guardrail Coverage Mapper
Map input, output, tool-input, tool-output, permission, and approval controls across an agent workflow.
Open tool →Agent Trace Coverage Scorecard
Check whether traces capture model calls, tools, handoffs, guardrails, latency, errors, identity, cost, and correlation IDs.
Open tool →Trace Span Analyzer
Paste trace spans and surface slow steps, failures, total duration, and the operations dominating latency.
Open tool →Agent Handoff Planner
Design agent-to-agent handoffs with trigger conditions, context transfer, ownership, validation, and fallback rules.
Open tool →Agent Eval Dataset Planner
Plan a balanced evaluation dataset across happy paths, edge cases, failures, safety cases, and historical production traces.
Open tool →Agent Regression Comparison Calculator
Compare baseline and candidate agent versions across success, tool accuracy, latency, cost, and safety failures.
Open tool →Context Compaction Planner
Estimate when a long-running agent should summarize, prune, externalize, or reload context.
Open tool →Session Memory Budget Calculator
Estimate how many turns fit inside a session budget after instructions, tools, retrieval, and output reserve.
Open tool →Agent Rate Limit & Concurrency Calculator
Estimate request throughput, required concurrency, token pressure, and safe queue capacity for an agent workload.
Open tool →LLM Cache Savings Calculator
Estimate token and cost savings when repeated prompt/context prefixes can be served from a lower-cost cache path.
Open tool →Agent Loop Detector
Inspect an agent trace for repeated calls, oscillating patterns, repeated failures, and runaway-step risk.
Open tool →Agent Run Replay Viewer
Turn raw trace JSON or line-based run logs into a readable step-by-step execution replay.
Open tool →Agent Failure Analyzer
Classify likely agent failure modes from traces and error text, then surface deterministic remediation checks.
Open tool →Failure to Eval Converter
Convert a failed production run into a reusable regression-test case with expected and forbidden behaviors.
Open tool →RAG Benchmark Calculator
Calculate Precision@K, Recall@K, Hit Rate, MRR, and nDCG from relevant IDs and a ranked retrieval result.
Open tool →Retrieval Result Inspector
Review ranked retrieval results with human relevance labels and summarize ranking quality without an LLM.
Open tool →RAG Experiment Comparator
Compare two retrieval configurations across recall, ranking quality, and latency using transparent weights.
Open tool →RRF Fusion Calculator
Combine dense and lexical rankings with Reciprocal Rank Fusion and inspect each document’s fused score.
Open tool →Parent-Child Chunking Designer
Plan hierarchical parent and child chunk sizes, overlap, estimated chunk counts, and expanded retrieval context.
Open tool →Agent Memory Architecture Builder
Choose working state, session memory, durable profile memory, episodic history, semantic knowledge, and retention policies.
Open tool →Tool Selection Benchmark Builder
Turn tool-routing scenarios into exportable benchmark cases for expected, forbidden, and no-tool decisions.
Open tool →Tool Contract Linter
Lint tool contracts for vague names, weak descriptions, schema mistakes, overlapping tools, and side-effect ambiguity.
Open tool →No tools found.
Try a broader term or switch back to All.