Ship AI tools without shipping your data.
AI Infrastructure
Sinas is the layer between your application and the LLM models. Your database stays where it is. The agents, functions and connectors run inside an instance that is yours, and you decide which provider ever sees what is in it, including one running on your own hardware.

Ship AI tools without shipping your data.
AI Infrastructure
Sinas is the layer between your application and the LLM models. Your database stays where it is. The agents, functions and connectors run inside an instance that is yours, and you decide which provider ever sees what is in it, including one running on your own hardware.

Ship AI tools without shipping your data.
AI Infrastructure
Sinas is the layer between your application and the LLM models. Your database stays where it is. The agents, functions and connectors run inside an instance that is yours, and you decide which provider ever sees what is in it, including one running on your own hardware.

Built so your
data stays yours.
Every AI feature sends something somewhere eventually. What differs is who decides what crosses, and whether you can see it afterwards. With Sinas that is configuration you control, not a policy you accept: your provider, your keys, your permissions, and nobody in your data.
Your provider, your key
Point an agent at OpenAI, Anthropic or Mistral using your own account. The contract with the model provider is yours, so their terms about your data are between you and them, with nobody proxying in the middle.
Your provider, your key
Point an agent at OpenAI, Anthropic or Mistral using your own account. The contract with the model provider is yours, so their terms about your data are between you and them, with nobody proxying in the middle.
Or no third party at all
Ollama is a provider like any other, so the model can run on hardware you own. For the work that can never leave the building, nothing has to.
Or no third party at all
Ollama is a provider like any other, so the model can run on hardware you own. For the work that can never leave the building, nothing has to.
Permissions decide what a model can see
An agent can only reach what the person using it can reach, and that is re-checked when a tool runs rather than when it is offered. What could be sent is bounded before anyone writes a prompt.
Permissions decide what a model can see
An agent can only reach what the person using it can reach, and that is re-checked when a tool runs rather than when it is offered. What could be sent is bounded before anyone writes a prompt.
We are not in your data
We do not read what is in your instance, mine it, or train anything on it. That sits in the terms rather than only on this page, and the platform is open source, so you can check the code that handles it.
We are not in your data
We do not read what is in your instance, mine it, or train anything on it. That sits in the terms rather than only on this page, and the platform is open source, so you can check the code that handles it.
Everything you need to ship AI applications.
AI Agents
Multi-provider LLM agents with tool calling, streaming, and agent-to-agent orchestration. OpenAI, Anthropic, Mistral, or Ollama.
AI Agents
Multi-provider LLM agents with tool calling, streaming, and agent-to-agent orchestration. OpenAI, Anthropic, Mistral, or Ollama.
AI Agents
Multi-provider LLM agents with tool calling, streaming, and agent-to-agent orchestration. OpenAI, Anthropic, Mistral, or Ollama.
Python Functions
Serverless Python in isolated Docker containers. Triggered by agents, webhooks, schedules, or data changes.
Python Functions
Serverless Python in isolated Docker containers. Triggered by agents, webhooks, schedules, or data changes.
Python Functions
Serverless Python in isolated Docker containers. Triggered by agents, webhooks, schedules, or data changes.
Connectors
Import an OpenAPI spec and Sinas generates typed operations your agents call as tools. Per-user OAuth, so one integration serves a whole tenant.
Connectors
Import an OpenAPI spec and Sinas generates typed operations your agents call as tools. Per-user OAuth, so one integration serves a whole tenant.
Connectors
Import an OpenAPI spec and Sinas generates typed operations your agents call as tools. Per-user OAuth, so one integration serves a whole tenant.
Components
Embeddable UI widgets compiled by Sinas. They reach agents and functions through a permission-enforcing proxy, so no credentials touch the browser.
Components
Embeddable UI widgets compiled by Sinas. They reach agents and functions through a permission-enforcing proxy, so no credentials touch the browser.
Components
Embeddable UI widgets compiled by Sinas. They reach agents and functions through a permission-enforcing proxy, so no credentials touch the browser.
Data & State
SQL templates against PostgreSQL, ClickHouse or Snowflake. Versioned file collections. Namespaced key-value memory with per-record encryption.
Data & State
SQL templates against PostgreSQL, ClickHouse or Snowflake. Versioned file collections. Namespaced key-value memory with per-record encryption.
Data & State
SQL templates against PostgreSQL, ClickHouse or Snowflake. Versioned file collections. Namespaced key-value memory with per-record encryption.
Triggers
Webhooks, cron schedules for functions or whole agents, and poll-based change data capture that needs nothing installed on the source database.
Triggers
Webhooks, cron schedules for functions or whole agents, and poll-based change data capture that needs nothing installed on the source database.
Who is Sinas for?
Two kinds of team, the same platform. What differs is what you are optimising for.
Developers and product teams
who need the AI feature shipped, not a platform to build first
Agents, functions, queries, state, files and skills behind one API, with permissions and logging already handled, so your team spends its time on the part users actually see.
Developers and product teams
who need the AI feature shipped, not a platform to build first
Agents, functions, queries, state, files and skills behind one API, with permissions and logging already handled, so your team spends its time on the part users actually see.
Developers and product teams
who need the AI feature shipped, not a platform to build first
Agents, functions, queries, state, files and skills behind one API, with permissions and logging already handled, so your team spends its time on the part users actually see.
Agencies and consultancies
who ship AI products for clients, one engagement after another
One isolated instance per client, a package library you reuse from the second project onwards, and a licence answer your client’s legal reviewer will accept.
Agencies and consultancies
who ship AI products for clients, one engagement after another
One isolated instance per client, a package library you reuse from the second project onwards, and a licence answer your client’s legal reviewer will accept.
Agencies and consultancies
who ship AI products for clients, one engagement after another
One isolated instance per client, a package library you reuse from the second project onwards, and a licence answer your client’s legal reviewer will accept.
USE CASES BY WEAREBRAIN
How WeAreBrain builds with Sinas.
Sinas is made by WeAreBrain, and our teams use it to build AI systems for clients. See real projects running on Sinas, from logistics to the public sector, on WeAreBrain.ai.
Logistics
Automating European truck breakdown operations from request to payment
Public sector
Building AI that keeps working when the internet does not
Consumer
Turning market signals into sales opportunities
USE CASES BY WEAREBRAIN
How WeAreBrain builds with Sinas.
Sinas is made by WeAreBrain, and our teams use it to build AI systems for clients. See real projects running on Sinas, from logistics to the public sector, on WeAreBrain.ai.
Logistics
Automating European truck breakdown operations from request to payment
Public sector
Building AI that keeps working when the internet does not
Consumer
Turning market signals into sales opportunities
USE CASES BY WEAREBRAIN
How WeAreBrain builds with Sinas.
Sinas is made by WeAreBrain, and our teams use it to build AI systems for clients. See real projects running on Sinas, from logistics to the public sector, on WeAreBrain.ai.
Logistics
Automating European truck breakdown operations from request to payment
Public sector
Building AI that keeps working when the internet does not
Consumer
Turning market signals into sales opportunities
How Sinas compares.
What each alternative is for, and when Sinas is the better fit.
01
Workflow automation
n8n · Zapier · Make
02
Coding harnesses
pi · Claude Code · Cursor
03
LangChain
LangChain · LangGraph
04
Visual builders
Dify · Flowise
05
Hosted agent APIs
OpenAI · Bedrock · Vertex AI
06
Build it yourself
Your codebase
VS
n8n
Zapier
Make
01 / 06
n8n and workflow automation
A workflow tool moves data between SaaS products when something happens.
Choose Sinas when
the thing you are building has users, permissions and data of its own, and not just two SaaS tools that need connecting.
VS
n8n
Zapier
Make
01 / 06
n8n and workflow automation
A workflow tool moves data between SaaS products when something happens.
Choose Sinas when
the thing you are building has users, permissions and data of its own, and not just two SaaS tools that need connecting.
Ownership, control, and a shorter road to working AI applications.
“The “no vendor lock-in” wasn’t just a marketing pitch. We have full RBAC and data sovereignty because it’s self-hosted. It’s rare to find a tool that offers this much power without forcing you into their cloud ecosystem.”
Maurice
Co-founder, TITAN
“We had a production-ready AI agent running in under 20 minutes. Being able to deploy the entire stack via Docker Compose meant we spent our afternoon building features instead of debugging environment variables.”
Dmitry Ermakov
CTO, Head of Engineering

“Finally, someone put agents, state, and files behind one clean API. It’s the abstraction layer we’ve been waiting for: clean, logical, and incredibly fast to implement.”
Enrico Karsten
CEO, Pulsr
Ready to start? You choose where it runs.
Sinas Cloud and the self-hosted edition are the same open-source platform, released under AGPL v3 and driven by the same configuration files. Start on Cloud in minutes with no card, or install it on your own infrastructure with one command. Move between the two whenever you like: it is an export and an import, never a rewrite.
WE HOST
Sinas Cloud
Managed for you · No card needed
A managed instance with upgrades, backups, monitoring and the console operated for you. Live in minutes, with no servers to size.
Your own instance, operated for you
Managed upgrades, monitoring and daily backups list
Every feature from the free tier up, nothing held back
Priority email support on every paid plan
WE HOST
Sinas Cloud
Managed for you · No card needed
A managed instance with upgrades, backups, monitoring and the console operated for you. Live in minutes, with no servers to size.
Your own instance, operated for you
Managed upgrades, monitoring and daily backups list
Every feature from the free tier up, nothing held back
Priority email support on every paid plan
YOU HOST
Self-hosted
AGPL v3 · One command to install
The same AGPL v3 platform on your own infrastructure. You own the data, the infra and the deployment.
One command on any Linux VPS: 2 cores, 4GB RAM, 50GB disk
Or the Helm chart on Kubernetes: kind, k3s, GKE or EKS
Airgapped, with password auth and no mail server at all
YOU HOST
Self-hosted
AGPL v3 · One command to install
The same AGPL v3 platform on your own infrastructure. You own the data, the infra and the deployment.
One command on any Linux VPS: 2 cores, 4GB RAM, 50GB disk
Or the Helm chart on Kubernetes: kind, k3s, GKE or EKS
Airgapped, with password auth and no mail server at all