Features
What Traverse does today: query languages, analytics, SDKs, deployment and operations. The docs cover each of these in depth.
Query Engine
Cypher 25
Traverse implements the Cypher 25 surface: quantified path patterns, dynamic labels and types, CALL IN TRANSACTIONS, COLLECT subqueries, vector similarity inside MATCH. If you know Cypher from Neo4j, you already know Traverse.
Full ISO GQL
100% conformance with the ISO/IEC 39075 GQL standard, including quantified path patterns, composed queries and catalog statements. The dialect is selectable per query, over any protocol and in Studio.
Automatic Parallelism
Heavy queries are automatically split across all available CPU cores. The query optimizer decides when parallelism helps - you just write queries.
Graph Algorithms & GDS
36+ algorithms built in and callable from Cypher: PageRank, Louvain, Dijkstra, similarity, embeddings and more, with stream, stats and write modes. The gds.* procedure surface works too, including named projections and mutate mode.
ML Pipelines
Train node-classification models, including graph neural networks, entirely from Cypher. GPU or CPU training, cross-validation, explainability, and a model catalog that persists with the database.
APOC Built In
The complete APOC core catalog, 246 functions and 190 procedures, ships in the binary: loaders, exports, triggers, background jobs, refactoring.
Deployment & SDKs
Server Mode
Run Traverse as a standalone service exposing Bolt, HTTP, and gRPC endpoints. Multi-database support, TLS encryption, API key authentication, and IP allowlisting.
Embedded Mode
Add Traverse as a library and run queries in-process - no network, no separate server. Available for .NET, Python, Java, Node.js, Go, and C/C++.
Bolt Protocol Compatibility
Full Bolt v5.x and v6.0 wire protocol support, so existing Neo4j drivers connect as they are. For clients that insist on a genuine Neo4j server, the advertised server agent is configurable.
Official HTTP & Embedded SDKs
First-class client libraries for .NET, Java, Python, Node.js, and Go. Each SDK supports HTTP, Bolt, or embedded connections with idiomatic APIs for your language.
Docker & Windows Service
An official Docker image, configured entirely through environment variables. On Windows, one command installs Traverse as a native service, and the binaries are Authenticode-signed.
In the Browser (WASM)
The full engine compiled to WebAssembly, with databases persisted in the browser. Try it at traverse.truespar.com, or embed it in your own app with the @truespar/traverse-wasm npm package.
Data Management
ACID Transactions
Every write is fully transactional - either it all succeeds or nothing changes. Readers never block writers and writers never block readers.
Durability, Your Choice
One portable .tvdb file per database, like SQLite, compressed and checksummed. Snapshot persistence by default; turn on write-ahead logging and every committed write survives a hard crash.
Multi-Database
Host multiple named databases on a single server. Create, drop, rename, and switch between databases - all via Cypher commands or Traverse Studio.
Database Aliases
Point multiple names to the same database. Useful for blue-green deployments - swap which database an alias points to without changing your application code.
Hot-Swap Reload
Unload a database from memory, replace the file, and reload - all without restarting the server. Zero-downtime database updates for production environments.
Read-Only Databases
Designate specific databases as read-only. Protect production data from accidental writes while still allowing full query access.
Indexes & Schema
Six Index Types
Range, composite and relationship-property indexes, plus text, point, BM25 full-text and HNSW vector indexes. The planner uses them automatically; USING INDEX hints exist when you want control.
Flexible Schema
No upfront schema required. Add labels, types, and properties as your data evolves - no migrations, no downtime. Structure your data however makes sense.
Constraints
Uniqueness, existence, composite NODE KEY and property-type constraints, on nodes and relationships. Definitions persist in the database file and participate in crash recovery.
Developer Tools
Traverse Studio
A web IDE bundled with every installation: query editor with autocomplete, graph visualization, schema browser, data import and database management. The GDS catalog and ML pipelines each have their own panel. Works on phones too.
CLI
Interactive command-line shell with syntax highlighting, tab completion, and query history. Connect to a server or open a database file directly. Output as table, JSON, or CSV.
MCP Server
Built-in Model Context Protocol support. AI assistants like Claude can query your graph, explore schema, find paths, and search nodes directly - no plugins or extra setup needed.
HTTP, gRPC & WebSocket APIs
50+ REST endpoints with interactive OpenAPI documentation. gRPC for high-throughput integrations. WebSocket for real-time streaming queries. All fully documented.
Bulk Import Tools
Import from CSV or Cypher files - via Studio, CLI, HTTP API, or the dedicated admin tool. Auto-indexing detects when your import switches from creating to querying and builds indexes automatically.
Graph Assistant
AI-powered assistant in Traverse Studio. Describe what you want in natural language and get Cypher queries generated from your schema. Available on Pro and Enterprise plans.
Operations & Security
Authentication & RBAC
User management with three roles - Admin, Editor, and Reader. Create and manage users via Cypher or the HTTP API. API key authentication for programmatic access.
TLS & IP Allowlisting
Encrypt all connections with TLS certificates. Restrict access to specific IP ranges with CIDR-based allowlisting.
Memory Management
Configurable memory budget with per-database tracking. Traverse enforces limits to prevent out-of-memory crashes and shows memory breakdown in Studio.
Query Timeouts & Monitoring
Set global query timeouts and slow query thresholds. Prometheus metrics endpoint for monitoring. Health and readiness probes for container orchestration.
Flexible Configuration
Configure via TOML file, CLI flags, or environment variables - with clear precedence rules. Auto-detects config files in the working directory.
File Management
Upload and download database files via the HTTP API. Move databases between machines by copying a single .tvdb file - no complex export/import procedures.
Feature Comparison
| Traverse | Neo4j | Memgraph | |
|---|---|---|---|
| Protocols & APIs | |||
| Bolt protocol | v5.x + v6.0 | v4.4 + v5.x + v6.0 | v1 - v5.2 |
| HTTP REST API | Yes (50+ endpoints) | No | No |
| WebSocket API | Yes | No | No |
| gRPC API | Yes | No | No |
| MCP server (built-in) | Yes | No | No |
| Interactive API documentation | Yes (OpenAPI + Scalar) | No | No |
| Query & Analytics | |||
| GQL (ISO/IEC 39075) | 100% conformance | Partial | No |
| APOC compatibility | Built in (246 functions, 190 procedures) | Plugin | No |
| Graph algorithm library | Built in (36+ algorithms + gds.* compat) | GDS (plugin) | MAGE (bundled) |
| ML pipelines & GNN training | Built in (GPU or CPU) | GDS ML (plugin) | No |
| Vector index | HNSW + quantization | Yes | Yes |
| Full-text index | BM25 | Lucene | Experimental |
| Triggers | Yes (apoc.trigger) | Via APOC plugin | Yes |
| Runs in the browser (WASM) | Yes | No | No |
| SDKs & Embedding | |||
| Embedded mode (in-process) | .NET, Python, Java, Node.js, Go, C | Java | No |
| .NET SDK | HTTP + Bolt + Embedded | Bolt | No (via Neo4j driver) |
| Java SDK | HTTP + Embedded | Bolt | No (via Neo4j driver) |
| Python SDK | HTTP + Embedded | Bolt | Bolt |
| Node.js SDK | HTTP + Embedded | Bolt | No (via Neo4j driver) |
| Go SDK | HTTP + Embedded | Bolt | No (via Neo4j driver) |
| Rust SDK | No | No | Yes (Bolt client) |
| C/C++ SDK | Embedded | No | Yes (Bolt client) |
| Database Operations | |||
| Multi-database | Yes (Free: 1, Pro: 10, Enterprise: unlimited) | Enterprise only | Enterprise only |
| Database aliases | Yes | Enterprise only | No |
| Hot-swap database reload | Yes | No | No |
| Read-only databases | Yes | Yes | No |
| Upload/download databases via API | Yes | No | No |
| Security | |||
| Role-based access control | Pro+ | Enterprise only | Enterprise only |
| API key authentication | Yes | No | No |
| IP allowlisting | Yes | No | No |
| Monitoring & Operations | |||
| Prometheus metrics | Yes | Enterprise only | Enterprise only |
| Per-database memory tracking | Yes | No | No |
| Query timeout configuration | Yes | Yes | Yes |
| Slow query logging | Yes | Yes | No |
| Platform & Deployment | |||
| Windows | Yes (+ Windows Service) | Yes (+ Windows Service) | No |
| Linux | Yes | Yes | Yes |
| macOS | Yes | Yes | No |
| Docker | Yes | Yes | Yes |
| In-memory engine | Yes | No (disk-based) | Yes |
| Parallel query execution | Yes | Enterprise only | Yes |
| Developer Experience | |||
| Visual IDE (bundled) | Traverse Studio | Neo4j Browser | Memgraph Lab |
| Graph visualization | Yes | Yes | Yes |
| Schema browser | Yes | Yes | Yes |
| Data import UI | Yes (Cypher + CSV) | Yes (Data Importer) | Yes (CSV) |
| AI assistant in IDE | Yes (Pro+) | Yes (Aura) | Yes (GraphChat) |
| Database upload & download | Yes | No | No |
| Saved queries & sharing | Yes | Favorites only | Yes |
| Extensibility | |||
| Custom query procedures | No | Yes (Java, APOC) | Yes (Python, C, C++, Rust) |
| Stream processing | No | No | Kafka, Pulsar, Redpanda |
Ready to try it?
Get started with the documentation or see detailed performance benchmarks.