SPACE PER CLIENT
Organize agents, channels and knowledge without mixing operations.
Build AI agents, connect your channels, and give every client a space of their own. Less repetitive work. More conversations that matter.
One space for your agents, your clients’ knowledge, and your team’s conversations.
Organize agents, channels and knowledge without mixing operations.
Connect AI providers and keep control of your usage.
Take over a conversation whenever you need to.
Deploy on your infrastructure under the FSL-1.1-MIT license.
Agents that know your business, tools that connect your processes, and a team that stays in control.
Set instructions, personality, per-client and per-agent context, timezone, and temperature / max-tokens / memory controls — then build in minutes with the creation wizard's live token counter and industry starter templates.
Upload PDFs, write structured Q&A pairs, or paste manual context. Retrieval is embedding-based semantic search with a keyword fallback — and embeddings are stored as portable JSON, no database extension required.
Give an agent per-agent HTTP tools — any REST endpoint, with path/query/body parameters, encrypted auth headers and a built-in SSRF guard — or connect MCP servers over Streamable HTTP or SSE, with test-before-save checks.
WhatsApp Cloud API with your own Meta app credentials, or WhatsApp QR through Baileys for a lighter setup — plus an embeddable web chat widget for any site. Instagram DM and Facebook Messenger are on the roadmap.
Connect your stack with a clear view of what is available and what needs setup.
Meta requires accounts, permissions and provider validation. Channel capabilities differ.
Explore channelsServer-side search, filter tabs, unread tracking and pagination across every channel — with one-click human takeover when a conversation needs a person.
A familiar presence while you configure agents and shape your operation.
Every client gets its own login and inbox, optionally served under their own custom domain — DNS-verified, with automatic HTTPS.
Your agency's name, identifier, color and logo throughout — plus a dashboard with activity, top agents and token usage by model.
—Scattered conversations
—Repeated setup per client
—Replies without context
—Errors without follow-up
An inbox for your operation
Client workspaces and agents
Knowledge connected to your agents
Visible attempts and send states
Run Voysse on your own server. Voysse does not meter seats or active clients; your infrastructure and providers set the practical limits.
Free, foreverGet startedA client counts as active once per month when its business has at least one real production conversation. Tests and playground activity do not count.
US$99 / monthFor agencies whose operational or security needs go beyond Cloud. We scope architecture, compliance, and support around your team.
Custom pricingThe software is FSL-1.1-MIT licensed, so Voysse does not charge you to self-host it; your own infrastructure and providers still have costs. Voysse Cloud is US$99/month for three active managed clients, plus US$25/month per additional active client, with AI and third-party channel charges separate.
OpenAI (Responses API) and Anthropic (Messages API) out of the box, agency-level and encrypted. You can also point any agent at any OpenAI-compatible endpoint via a per-connection base URL and model.
Yes — Voysse is self-hosted. Three services (API, web, WhatsApp bridge) plus PostgreSQL run behind one Docker Compose file and a Caddy gateway, so it's one command on any server you control.
Every query is scoped by agency_id at the data layer. AI provider keys and WhatsApp session state are encrypted at rest, and public endpoints are rate-limited per client IP.
WhatsApp Cloud API, WhatsApp via QR (Baileys), and an embeddable web chat widget for any site. Instagram DM and Facebook Messenger are planned next.
Give your agency a space to grow with AI.