The facts, with their sources

What it does

Done-for-you service: YesBlits.ai sells a delivered use case: its team workshops, builds, integrates, deploys and then runs the agent for its whole life.You buy a working use case, not a license to build one. The Blits.ai team takes it from workshop to production and runs it for as long as it is live.blits.ai/products/agentic-ai

How much it does unattended

Runs autonomously: PartlyWorkflow runs start and finish without a person, but tools that move money or change data wait for approval by default.A workflow starts from a schedule, a webhook, an incoming email, a new file, an API call or a manual run. It then works until the goal is reached, a…blits.ai/how-agents-work
Multi-agent: YesAgents delegate sub tasks to specialist agents and multi agent workflows run on shared context and shared goals.Run multi agent workflows on shared context and shared goals.blits.ai/products/agentic-ai
Agent permissions: YesA per tool execution policy decides automatic, approval-first or forbidden, with read-only tools running alone and money-moving tools gated.Tools that move money or change data require approval by default, and read only tools run automatically. Each tool can also be set to automatic or…blits.ai/security-and-governance

Which models it runs on

Claude: YesAnthropic Claude is named among the 300-plus connectable models, and the model can be chosen or switched per agent.Connect to more than 300 models from OpenAI, Azure OpenAI, Anthropic Claude, Google Gemini and Gemma, Meta Llama, Mistral, Microsoft Phi, Alibaba…blits.ai/features
GPT: YesOpenAI and Azure OpenAI models are both listed as providers you can connect and select per agent.Connect to more than 300 models from OpenAI, Azure OpenAI, Anthropic Claude, Google Gemini and Gemma, Meta Llama, Mistral, Microsoft Phi, Alibaba…blits.ai/features
Open models: YesOpen-weight families are supported and can run inside the buyer's own environment, including fully air gapped deployments.For strict residency, choose models that a cloud provider hosts in the same region, or open weight models that run inside your own environment,…blits.ai/data-residency
Model choice: YesThe buyer picks the model per agent and per task and can switch providers without rebuilding the workflow.Combine Google, OpenAI, Claude, and more on Blits.ai. Switch providers whenever you choose, without rebuilding a single workflow.blits.ai/products/agentic-ai

Where you use it

In your editor: NoThe vendor's complete 190-feature inventory lists fourteen channels covering chat, voice, Teams and APIs, and no IDE or code-editor extension among them.A complete, searchable overview of the Blits.ai platform. From AI agents and channels to payments, guardrails, and enterprise governance, all in one…blits.ai/features
On the command line: NoNo terminal or command-line tool appears in the complete feature inventory; programmatic use is via REST and WebSocket APIs instead.A complete, searchable overview of the Blits.ai platform. From AI agents and channels to payments, guardrails, and enterprise governance, all in one…blits.ai/features
In your pipeline: PartlyRelease gates promote an assistant only when a chosen test suite passes, but this runs inside the platform, not a buyer's CI on pull requests.Promote an assistant to production only when a selected test suite, including latency checks, passes on the new version.blits.ai/features
In a browser: YesThe platform is a hosted web application at platform.blits.ai with builders, dashboards and a tenant admin console.No code and pro code builders, so your teams iterate fastblits.ai/products/agentic-ai

Whose machine it runs on

Self-hosted: YesThe Kubernetes-native platform runs in the buyer's own cloud account or data centre, including air gapped installs done on site.Run entirely in your own data center for maximum control and security. Built for organizations with strict data residency requirements.blits.ai/deployment-options

What it costs to run

How it meters: YesPlans meter monthly interactions alongside bot and seat counts, with extra interactions and AI engine costs listed as separate lines.1k interactions/month 1 bot 1 seat 2 channels 2 AI engines AI engine costs Hosting of bots Self guided onboarding Extra interactionsblits.ai/pricing
Free tier: NoAll four plans read Contact us, prices are on request and billed annually in advance; the only no-cost route is a sales demo.Prices are available on request. All prices exclude VAT and are charged annually in advance.blits.ai/pricing
API access: YesREST and WebSocket APIs start conversations, stream answers and run workflows, with API tokens per agent, bot and workflow.Start conversations, send messages and stream answers over REST or WebSocket, and run workflows through the API.blits.ai/integration-standards
MCP server: YesIt is both MCP client and MCP server: the platform hosts an API-key protected MCP server per app so other agents can call its tools.The platform hosts an MCP server per app, protected by an API key, that gives other agents and assistants tools over the data you store in the…blits.ai/integration-standards
Bring your own key: YesBuyers set their own key and endpoint per agent for any OpenAI-compatible model, and usage is tagged platform key or customer key.Set your own API key and endpoint per agent. Usage is tagged as platform key or customer key, so costs stay traceable.blits.ai/integration-standards

Buying it for a team

A company can buy it: YesEvery plan is bought through sales on annual invoicing, with multi-tenant hierarchy, subtenants and an admin console for the organisation.Four plans on one platform. Start with a single bot, scale to unlimited bots, and get pricing built around your volume and requirements.blits.ai/pricing
Seat model: YesSeats are capped per plan rather than sold individually: 1 on Basic, 5 on Pro, 15 on Business, unlimited on Enterprise.15 seatsblits.ai/pricing
Pooled budget: YesEach plan carries one monthly interaction allowance for the whole tenant, with extra interactions billed beyond it, so consumption is shared not per seat.250k interactions/monthblits.ai/pricing
Admin controls: YesA tool execution policy sets per tool whether an agent may run it, must ask first, or may not use it at all.Set per tool whether an agent may run it automatically, needs approval first, or may not use it at all.blits.ai/features
Audit log: YesUser actions and sign ins are logged and exportable as CSV, and every workflow run stores its tool calls and returns.Changes and sign ins by platform users are logged, and tenant admins can export the audit log as CSV.blits.ai/security-and-governance
Single sign-on: PartlyMicrosoft Entra and Google SSO are supported for admin login; the vendor names no SAML or generic OIDC provider support.Single sign on Microsoft (Azure AD / Entra) and Google SSO.blits.ai/features

What happens to your code

Opt out of training: YesThe privacy policy rules out training on customer data by default and says model providers are contracted on terms that forbid it too.Blits.ai does not use customer data or conversation content to train, fine tune or improve any artificial intelligence model, whether our own or that…blits.ai/privacy-policy
Data residency: YesSix named managed regions each hold a tenant's data in place, and any other country is served by private cloud or on premises.Blits.ai runs managed production environments in the European Union, Qatar, Singapore, the United Arab Emirates, the United States and Australia.blits.ai/data-residency
Getting out: YesAssistants export as one encrypted package with BPMN flows, and conversations, training data and consent records export as CSV.An assistant exports as one package with its flows, intents and training data, functions, and the agents, workflows and model settings it uses. Flows…blits.ai/data-portability
Certifications: YesFour independent certifications are claimed for the platform and managed service, with GDPR and EU AI Act alignment alongside them.ISO 27001, ISO 42001, SOC 2 Type II and PCI DSS Level 1.blits.ai/security-and-governance