One Platform, Limitless Capabilities
8 integrated apps. 233 AI agents. 363+ tools. 14 microservices. ChaozCode is the AI DevOps platform that gives your team superpowers.
Memory Spine is a core component of the ChaozCode AI Development Platform.
Memory Spine — Persistent AI Memory
The foundation of ChaozCode’s intelligence. Memory Spine gives every agent, app, and workflow persistent recall — so context is never lost between sessions, handoffs, or deployments.
Core Memory Operations
Store, search, retrieve, update, delete, and pin memories. Every operation is vectorized automatically. Pin critical system configs for instant recall across all agents.
memory_store · memory_search · memory_retrieve · memory_pin
Hybrid Search — Sub-25ms
FTS5 full-text search combined with vector similarity in a single query. Ranked results with relevance scoring. 250K+ vectors at sub-25ms average latency.
FTS5 + vectors · Query expansion · Relevance scoring
Knowledge Graphs
Auto-build knowledge graphs from memories. Extract entities, map relationships, traverse connections. Query the graph for patterns your team didn’t know existed.
knowledge_graph_build · knowledge_graph_query
Conversation Intelligence
Track multi-turn conversations with context management. Start sessions, add turns, get token-optimized context windows, and save conversations to long-term memory.
conversation_start · conversation_save · conversation_get_context
Codebase Context
Pre-edit context loading with relevant file analysis and dependency mapping. Deep repository analysis with complexity metrics and architecture insights.
codebase_context · codebase_analyze · codebase_suggest
Multi-Agent Handoff
Seamless context transfer between specialized agents. Include recent memories, specify targets, and maintain continuity across agent boundaries.
agent_handoff · include_recent · target_agent
LLM Context Window
Build optimized LLM context from relevant memories with configurable token budgets. Maximize signal, minimize waste. Up to 32K tokens per context window.
llm_context_window · token budget · relevance ranking
Batch Operations
Bulk store and bulk tag hundreds of memories in a single call. Efficient categorization workflows for high-throughput pipelines and migration scenarios.
memory_batch_store · memory_batch_tag
Memory Spine — Store & Search
// Store a memory with automatic vectorization
await memory_store({
content: "Deployed auth service v2.3 with OAuth2 PKCE flow",
tags: ["deploy", "auth", "v2.3"],
metadata: { service: "auth", version: "2.3.0" }
});
// Hybrid search — FTS5 + semantic vectors in one call
const results = await memory_search({
query: "authentication deployment issues",
limit: 10
});
// Results ranked by combined FTS5 + vector relevance
results.forEach(r => {
console.log("[" + r.score + "] " + r.content);
console.log(" Tags: " + r.tags.join(", "));
});
Agent Intelligence — AgentZ + ML Router
233 specialized agents. Automatic task routing. Parallel execution. Every agent is an expert in its domain — and the ML Router learns which agent handles each task best.
233 Specialized Agents
Agents for every task: explore, code-review, security-auditor, performance-tuner, auto-fixer, test-specialist, refactorer, architecture analyzer, and 225 more. Each built with domain-specific knowledge.
explore · code-review · security-auditor · auto-fixer · 229 more
ML Router — Auto-Routing
Describe a task in plain English. The ML Router analyzes intent, matches capabilities, and routes to the best agent automatically. Learns from feedback to improve over time.
mlrouter_route_task · mlrouter_suggest_agents · mlrouter_feedback
Parallel Swarms
Launch 3+ agents simultaneously on different aspects of a task. Code review, testing, and security audit — all running in parallel, results merged at completion.
Parallel execution · Result aggregation · No conflicts
Sequential Pipelines
Chain agents for multi-step workflows: explore, fix, test, review. Each agent receives the output of the previous. Build complex automations from simple agent calls.
Explore → Auto-fix → Test → Code-review
Agent Handoff via Memory
Transfer full context between agents through Memory Spine. Include recent memories, conversation history, and task state. Zero context loss between handoffs.
agent_handoff · include_recent: 20 · target_agent
Cost-Gated Execution
4-tier cost model: Trivial ($0.001), Routine ($0.01), Standard ($0.05), Complex ($0.20). Platform automatically selects the cheapest agent that handles the task. Opus for complex reasoning only.
Haiku → Grok → Sonnet → Opus · Auto-tier selection
Agent Routing — ML Router
// Auto-route a task to the best agent
const route = await mlrouter_route_task({
description: "Review the auth module for SQL injection vulnerabilities"
});
// Routes to: security-auditor (confidence: 0.94)
// Or get top 3 agent suggestions
const suggestions = await mlrouter_suggest_agents({
description: "Optimize database queries in the user service",
top_n: 3
});
// [performance-tuner: 0.91, query-optimizer: 0.88, database-expert: 0.85]
// Parallel swarm — 3 agents at once
await Promise.all([
task({ agent: "code-review", prompt: "review auth module" }),
task({ agent: "security-auditor", prompt: "audit auth module" }),
task({ agent: "test-specialist", prompt: "write tests for auth" })
]);
Code Intelligence — Zearch + HelixHyper
Find anything in any codebase. Understand how code connects. Zearch searches by meaning, HelixHyper maps relationships. Together, they give you x-ray vision into your code.
Zearch Deep Search
623 endpoints covering 60+ file types. Full-text, regex, and semantic search modes. Find code by what it does, not just what it’s called. Cross-file dependency tracing.
623 endpoints · 60+ file types · Semantic + regex + FTS
Cross-Repo Analysis
Search across multiple repositories in a single query. Find shared patterns, duplicated code, and cross-project dependencies. Unified results with repository attribution.
Multi-repo search · Dependency tracing · Deduplication
Knowledge Graph Engine
811 nodes with PageRank, degree centrality, betweenness centrality, and eigenvector centrality. Identify the most critical code paths and architectural bottlenecks.
PageRank · Betweenness · Eigenvector · 811 nodes
Community Detection
Automatically find code clusters, module boundaries, and dependency groups. Detect tightly-coupled components and suggest decomposition strategies.
Cluster analysis · Module boundaries · Coupling metrics
Pattern Analysis
Detect structural patterns: triangles (circular dependencies), stars (hub modules), bridges (critical connectors), and isolated nodes (dead code candidates).
Triangles · Stars · Bridges · Isolated nodes
Semantic Code Search
Search code by meaning, not keywords. Ask “functions that validate user input” and get results ranked by semantic relevance. Understands intent across languages.
Natural language queries · Cross-language · Intent matching
Developer Experience — ChaozPilot + GitChaozOxide
Ship faster. ChaozPilot gives you real-time AI assistance while coding. GitChaozOxide handles the git intelligence — blame, PRs, security scanning, and branch management.
ChaozPilot IDE Integration
Real-time code completion powered by your full codebase context. Memory Spine feeds project history, Zearch provides relevant code snippets, agents supply domain expertise.
Real-time · Context-aware · Memory-backed completions
Inline Suggestions
Context-aware code suggestions that understand your project’s patterns, conventions, and architecture. Not just autocomplete — intelligent code generation inline.
Pattern-aware · Project conventions · Inline generation
Git Intelligence
Blame analysis, commit history tracking, and diff comparison built in Rust for speed. Understand who changed what, when, and why — across your entire repository.
git-blame · git-history · git-diff · Rust-powered
PR Automation
Automated code reviews with security scanning. Create, review, and merge pull requests programmatically. AI-powered merge conflict resolution.
Auto-review · Merge resolution · Security scan on push
Security Scanning
Detect secrets, API keys, credentials, and vulnerabilities before they reach the remote. GitChaozOxide scans every commit with configurable depth.
Secret detection · Credential scanning · Pre-push hooks
Branch Management
Create, switch, list, and delete branches. Full lifecycle management with conflict detection and auto-merge capabilities. Works with GitFlow, trunk-based, and custom strategies.
Create · Switch · Delete · Lifecycle management
Technical Specifications
Platform
8 Integrated Apps
Total Agents
233 Specialized
Total Tools
363+
Microservices
14 Independent
API Protocol
REST + MCP
Auth
JWT + API Keys + Service Tokens
Search Latency
Sub-25ms
Vector Capacity
250K+
Every feature included in every ChaozCode plan
Starter
1K vectors · 5 agents/day
$0/mo
POPULAR
Indie
10K vectors · 25 agents/day
$29/mo
Developer
50K vectors · 100 agents/day
$49/mo
Master
250K vectors · Unlimited agents
$99/mo
Plus Business ($249), Business+ ($599), Business Large ($999), and Enterprise plans
Compare all 8 plans →