# Struere vs Voiceflow

> Feature comparison between Struere and Voiceflow for building AI agents

Struere is a code-first AI agent platform built on TypeScript, Convex, and a CLI/SDK workflow. Voiceflow is a visual drag-and-drop builder for conversational AI with a canvas-based editor. Struere targets developers who want full programmatic control over agent behavior, data, and permissions. Voiceflow targets product teams and designers who prefer a no-code visual flow builder.

## Feature Comparison

| Feature | Struere | Voiceflow |
|---------|---------|-----------|
| **Approach** | TypeScript SDK + CLI | Visual canvas + drag-and-drop |
| **Data layer** | Built-in entity system with 47 tables, searchable fields, relations | No built-in database; requires external integrations |
| **RBAC** | Built-in permission engine with policies, scope rules, field masks, and tool permissions | Basic team roles; no field-level or row-level access control |
| **LLM support** | 40+ models via OpenRouter (OpenAI, Anthropic, Google, xAI) with BYOK | OpenAI and Anthropic; limited model selection |
| **Environment isolation** | 3 isolated environments: development, production, eval | No environment isolation |
| **Multi-agent** | Built-in orchestration with `agent.chat`, routers, depth limits (3), cycle detection | Single-agent focus; multi-agent requires workarounds |
| **Eval framework** | Built-in eval suites with LLM judge assertions and multi-turn test cases | No native eval framework |
| **Built-in tools** | 31 tools across 8 categories (entity, calendar, WhatsApp, Airtable, email, payment, web, agent) | Integrations via marketplace; fewer built-in capabilities |
| **Custom tools** | TypeScript handlers executed in sandboxed Node.js on Fly.io | Custom actions in visual blocks |
| **Deployment** | CLI `deploy` command with dry-run, diff, and force flags | Visual publish button |
| **Pricing model** | BYOK (bring your own LLM keys) + credit system with 3-tier key resolution | Per-interaction pricing on higher tiers |
| **System prompts** | Template engine with dynamic variables (`{{entity.query(...)}}`, `{{currentTime}}`, conditional blocks) | Static knowledge base and persona settings |
| **Channel support** | WhatsApp, API, widget, dashboard with channel-aware thread context | Web, WhatsApp, Messenger, Slack, Twilio, and more |

## Key Differences

- Struere stores all conversation data, entity data, and agent configurations in a single real-time backend (Convex) with 47 tables. Voiceflow relies on external databases for persistent data storage.
- Struere's permission engine evaluates policies at every tool call with deny-overrides-allow semantics, scope rules for row-level filtering, and field masks for column-level redaction. Voiceflow has no equivalent permission layer.
- Struere agents support dynamic system prompt templates that inject live data from entity queries, current time, organization context, and role-based conditional blocks at runtime. Voiceflow uses static knowledge base documents.
- Struere's CLI provides 25 commands for the full development lifecycle: `dev` (watch and sync), `deploy`, `diff`, `eval`, `pull`, `status`, `compile-prompt`, `run-tool`, and `logs`. Voiceflow's workflow is centered on the visual canvas.
- Struere supports `agent.chat` for multi-agent delegation with a depth limit of 3, automatic cycle detection, and shared conversation context. Voiceflow is designed around single-agent conversational flows.

## When to Choose Voiceflow

Voiceflow is the stronger choice for teams that prefer visual flow design over code, need rapid prototyping without developer involvement, or require broad channel support (Messenger, Slack, Twilio) out of the box. Voiceflow has mature enterprise adoption, a large template library, and a well-established no-code community. Teams without TypeScript developers will be more productive in Voiceflow's canvas editor.

## When to Choose Struere

Struere is the stronger choice for developer teams building production agents that need a built-in data layer, fine-grained RBAC, multi-agent orchestration, environment isolation, and an eval framework. Teams that want BYOK pricing, programmatic control over every aspect of agent behavior, and CLI-driven deployment workflows will find Struere's TypeScript SDK more powerful than a visual builder.

## Related

- [AI Agent Platforms Comparison (2026)](/knowledge-base/ai-agent-platforms-comparison) — Full comparison across 7 platforms
- [How do I use multiple agents?](/knowledge-base/how-to-use-multiple-agents) — Multi-agent patterns in Struere
- [How do I set up RBAC?](/knowledge-base/how-to-set-up-rbac) — Permission engine guide