Guides
Guides
Comparison guides and decision frameworks for AI automation. Each one leads with a verdict, then shows the numbers, the tradeoffs and the edge cases that change the answer.
October 24, 2026 · 5 minBest LLM for Document Extraction: A 50-Document Test (2026)
The best LLM for document extraction is the one that wins on your documents: schema-enforced output, field-level scoring, 50 test files, code.
October 24, 2026 · 5 minFalkorDB vs Neo4j: Which Graph Database for Agent Memory
FalkorDB vs Neo4j for GraphRAG and agent memory: SSPL or GPLv3, one graph per tenant or one database, and which one Graphiti now defaults to.
October 24, 2026 · 5 minZep vs Graphiti: Managed Memory or Self-Hosted Graph (2026)
Zep vs Graphiti: one temporal graph engine, two owners. Zep Cloud from $125 a month, the EU DPA catch, and when self-hosting pays off.
October 24, 2026 · 5 minGraphiti MCP Server: Setup for Claude Code, Cursor and Desktop
Graphiti MCP server in one container: connect Claude Code, Cursor or Claude Desktop, set group_id and stop the 429s that break ingestion.
GuideGraphiti Without Neo4j September 23, 2026 · 5 minGraphiti Without Neo4j: One Embedded File, Ollama, No API Key
Graphiti without Neo4j or Docker: Kuzu is deprecated, FalkorDB Lite allows one process. An embedded file with one writer and many readers, measured.
GuideGraphiti add_episode September 18, 2026 · 8 minGraphiti add_episode: Parameters, Indices and the Errors I Hit
How to call Graphiti's add_episode: parameters, EpisodeType, build_indices_and_constraints, group_id on FalkorDB, and the errors my graphs hit.
September 14, 2026 · 7 minPalantir Foundry Alternative: Start With One Workflow
Compare Palantir Foundry with a focused open-source context layer: scope, costs, governance and a runnable delivery-blocker demo.
GuideTemporal Knowledge Graph September 11, 2026 · 11 minTemporal Knowledge Graph: How AI Agents Know What Is True Now
What a temporal knowledge graph is, how valid time and transaction time work, and a four-step test of whether your AI agents need one or retrieval is enough.
GuideGraphiti vs Mem0 September 11, 2026 · 9 minGraphiti vs Mem0: Which Memory Layer Fits Your AI Agents
Graphiti vs Mem0: compare changed facts, access controls and operating cost. Use four acceptance checks before choosing your agent memory layer.
GuideGraphiti in Production September 11, 2026 · 7 minGraphiti in Production: FalkorDB or Neo4j, Azure or Ollama, and What Breaks
Graphiti in production on FalkorDB or Neo4j, from graphs I run: one embedding model on two Ollama builds agreed at cosine 0.80. The checklist that followed.
July 5, 2026 · 12 minClaude API Pricing Tiers and Cost Optimization Playbook (2026)
Claude API and Fable 5 pricing, plus every lever to cut Anthropic API cost without losing quality: routing, caching, batching, effort tuning.
July 5, 2026 · 9 minClaude Fable 5 Cost: What It Actually Costs and How to Control It (2026)
Claude Fable 5 and 5.1 cost $10/$50 per million tokens, twice Opus 5. Cache reads fell to $0.25 on 5.1. What drives the bill and how to govern it.
Your agents answer from whatever the retriever finds, and too often that is last quarter's truth. I build the context layer they answer and act from: a temporal knowledge graph that keeps every fact with its source and the time it held, reads with each person's own permissions, and writes nothing without a person's approval. On your own tenant, billed by the hour, step by step.