Head-to-head

SyntheticBrew vs Botpress: visual studio, or an agent in your product?

Botpress is a polished cloud studio and helpdesk billed per conversation, with model spend bundled into the quota. SyntheticBrew is an agent runtime you embed in your own product — priced by active users, paying token prices on your own LLM keys, with setup driven from Cursor or Claude Code via native MCP. All Botpress figures below are from their public pricing page, July 2026.

Side by side

SyntheticBrew vs Botpress: side-by-side comparison

Dimension	SyntheticBrew	Botpress
What it is	An agent runtime you embed in your own product — website widget plus a REST and SSE API	Cloud platform: AI-native helpdesk (Desk) plus a visual Agent Studio, managed on Botpress infrastructure
Pricing model	Priced per monthly active user; with your own key, messages are unlimited on every plan	Per-conversation: $150/mo for 250 conversations (Plus, billed annually), overage packs at $0.50–0.65 per conversation; LLM spend is bundled into the conversation quota
Overage control	No metered conversations — nothing to auto-recharge and nothing to expire	Auto-recharge cannot be turned off (their FAQ); unused pack conversations expire at month end
Who pays for inference	You do, directly — bring your own provider key and pay list price with zero markup	Bundled into the conversation quota, so model spend and platform spend cannot be separated or shopped
Bring your own LLM key	Yes, on every tier including free — add models on your own provider keys (any OpenAI-compatible endpoint or OpenRouter); plus an optional per-request end-user BYOK mode admins can enable	Documented, but Enterprise-only: their SDK docs route "use your own LLM" through a sales conversation. On self-serve plans LLM spend is bundled into per-conversation pricing
Manage from a coding agent (MCP)	Hosted MCP server in the official registry: connect Cursor or Claude Code over OAuth and provision agents, knowledge, and the embed snippet — no local install	ADK dev tooling ships a local MCP server (`adk mcp`) so coding assistants can work with your running dev project
Build style	Config-as-code agents over REST — or let your coding agent set everything up via MCP; admin dashboard included	Visual flow builder (Studio) with an agent SDK (ADK); polished no-code/low-code experience
Built-in channels	Website widget + REST/SSE API today; other channels bridge through your backend	Webchat and WhatsApp on all plans; webchat white-labelling from Plus; voice on Enterprise only
Knowledge grounding	Knowledge bases with RAG plus typed knowledge graphs; grounded answers with source attribution	Knowledge bases with vector storage (100 MB free / 1 GB Plus / 5 GB Team), tables for structured data
Data portability	Agents, knowledge, and session history are readable and exportable through the API; agent config lives as code in your repo	Vendor-managed cloud; storage capped by tier with a $40/mo expansion add-on

Where SyntheticBrew differs

Where SyntheticBrew differs from Botpress — and why it matters

Your model choice, and your inference bill

Botpress bundles LLM spend into the conversation quota, so you cannot separate model cost from platform cost, shop providers on price, or switch a single agent to a cheaper model. SyntheticBrew runs on any OpenAI-compatible endpoint or OpenRouter using your own key, switchable per agent, and never marks up a token: inference is billed to you by your provider at list price.

Why it matters: Model prices move every few months and quality varies by task. When inference is bundled, every one of those improvements accrues to the vendor instead of you — and you have no lever to pull when a support workload gets expensive.

Token prices, not conversation arithmetic

Botpress bills per conversation with LLM spend bundled in: $150/month covers 250 conversations, overages cost $0.50–0.65 each, auto-recharge cannot be disabled, and unused pack conversations expire monthly. SyntheticBrew has no conversation meter — you are billed by how many distinct people your agent talks to, and you pay your LLM provider’s token price directly.

Why it matters: A busy support widget can burn through 250 conversations in days. When each extra exchange costs half a dollar and top-ups are mandatory, cost scales with traffic — not with the value the agent delivers.

A hosted MCP, not just dev tooling

Both platforms speak MCP — differently. Botpress’s ADK ships a local MCP server so coding assistants can work on your running dev project. SyntheticBrew’s MCP server is hosted and listed in the official registry: your coding agent connects over OAuth and provisions the production agent, loads knowledge, and returns the embed snippet — one prompt, nothing installed locally.

Why it matters: Dev-loop tooling helps you build; a hosted MCP surface means the whole deployment — from empty account to a widget on your site — is driven from your editor and reproducible on demand.

For enterprise teams

An enterprise-ready Botpress alternative

Bring your own model keys — per request or per agent, used once and never logged.
Ed25519-signed JWTs, scoped API tokens, and an audit trail of agent actions.
No per-conversation metering at any volume — costs scale with reach, not chattiness.
Confirmation gates before consequential actions, with the result verified after the write.

Enterprise AI agents

Choose for the job

Pick by what you value

Want a managed no-code studio with WhatsApp and a built-in helpdesk? Botpress Cloud does that today. Want predictable token-price economics on your own model keys, an API from the free plan, answers grounded in your live systems, and agents provisioned by your coding agent? That is what SyntheticBrew is built for — and you can prove it on your own data before you decide.

Try Cloud Connect your editor Contact sales

Questions

SyntheticBrew vs Botpress: FAQ

Is SyntheticBrew a good Botpress alternative?

If you want token-price economics instead of per-conversation billing, your own choice of model, or setup driven from Cursor / Claude Code via MCP — yes. If your team prefers building in a polished visual studio with built-in WhatsApp and a managed helpdesk, Botpress Cloud serves that profile well.

Can I use my own LLM key with a Botpress alternative?

Botpress bundles LLM spend into its conversation quota, so there is no separate provider key to bring and no per-agent model shopping. SyntheticBrew is bring-your-own-key by default: add any OpenAI-compatible endpoint or OpenRouter, switch models per agent, and pay your provider directly at list price with no markup from us.

What does each actually cost to run?

SyntheticBrew is $0 to start and priced by monthly active users, with unlimited messages on your own key. Botpress Plus is $150/month (billed annually) for 250 conversations with LLM spend bundled; extra conversations cost $0.50–0.65 in packs of 100, auto-recharge cannot be turned off, and unused pack conversations expire at month end.

When is Botpress the better choice?

When you want a managed, no-code experience: a visual flow studio, a built-in helpdesk with routing and RBAC (Team plan), WhatsApp out of the box, and you are comfortable with cloud-only hosting and per-conversation pricing. Test both against your actual traffic profile.

Can I migrate from Botpress to SyntheticBrew?

Your knowledge sources are your own docs and site content — re-ingesting them into a SyntheticBrew knowledge base takes minutes, from the dashboard or via MCP from your editor. Flow logic built in Botpress Studio would be re-expressed as agent instructions and tools rather than imported.

More comparisons: SyntheticBrew vs Chatbase · SyntheticBrew vs Flowise · SyntheticBrew vs Dify · All comparisons