Tines 3B is an AI-native business workflow platform released in July 2026 by Tines, a workflow automation company based in Ireland. It lets you build, run, and monitor AI agents, internal apps, and automated workflows in a single environment, starting from either a natural-language prompt or code. The goal is to bring the unidentified business code that employees create with AI (a phenomenon Tines calls “wild code”) under the oversight of IT and security teams. It is a new product built separately from the Tines known for security automation (SOAR); the existing product has been renamed “Tines Stories” and continues to be developed in parallel.
Key Features
- Build from prompts or from code: You can start by describing the task you want in plain language, while the output is treated as real code rather than a no-code abstraction layer. Git integration and native branching come standard, so both human and AI changes can be tracked and reviewed
- Per-step isolated execution with hidden credentials: Each step runs as its own independent computer, fully separated from the next. Credentials are injected into the runtime through a transparent proxy, so they are never visible to the builder or the AI and never stored in code
- Built-in mechanisms to fix things before they break: During construction, Tines 3B suggests tests and generates dummy data, and asks for confirmation before running anything that touches production data or apps. Once live, Autofix detects anomalies and fragile spots and proposes fixes, while Autotune suggests reliability, performance, and readability improvements as reviewable branches
- Governance through Spaces, Skills, and RBAC: You carve out “Spaces” with connectors and permissions scoped to each use case, and use “Skills” to standardize LLM behavior for consistency across the organization. Custom roles and SSO are also supported
- A dashboard for measuring cost and value: AI spend by LLM provider and model, estimated time saved per workflow, duplicate workflow detection, and a list of dependency packages can all be monitored on a single screen
- Operate from external AI clients via MCP: A remote MCP endpoint lets you connect from tools such as Claude Code. Permissions depend on the authentication method: OAuth sign-in allows creating and publishing, while a service account token allows viewing and running
Pricing
Rather than buying features individually, you license the whole platform. Plans are differentiated by infrastructure configuration and execution capacity.
| Plan | Positioning | Execution capacity (Tines units / month) |
|---|---|---|
| Explore | Free trial edition | 5,000 TU (up to 3 active workflows, a one-time $50 AI usage credit, no public URLs) |
| Core | Multi-tenant configuration. For small to mid-sized teams | 150,000 TU |
| Scale | Dedicated stack. For large-scale, performance-focused use | 300,000 TU |
| Power | Premium dedicated stack. For the largest deployments | 600,000 TU |
- What a TU is: A proprietary execution metric that combines compute, storage, and data transfer
- No rollover: Resets monthly, with overages billed separately
- LLM costs are separate: Bring-Your-Own-Model (BYOM), so you use models under your own contracts
- Prices are not published: Paid plans are quote-based
Pricing information is as of August 2026. For the latest pricing, check the official pricing page and the documentation.
Pros and Cons
✅ Pros
- Consolidates the automations that tend to sprawl across a company, including those built by non-engineers, into one environment that IT and security teams can see in full
- Credentials are never exposed to the builder or the AI, and execution is isolated per step, so the blast radius of runaway generated code stays contained
- Visualizes AI spend by model, directly addressing the common complaint that generative AI costs are impossible to predict
- Existing Tines customers can try it at no extra cost, so you can evaluate it in parallel without shutting down Tines Stories
⚠️ Cons
- The intended audience is corporate IT and security departments, so it can easily be overkill for individuals or small teams
- Paid plan pricing is not public, so estimating adoption costs requires a sales conversation
- Because of the BYOM model, you need a separate contract with an LLM provider and must manage those costs yourself
- Having launched only in July 2026, external case studies and third-party reviews are still scarce
How It Differs from Tines Stories (the Original Tines)
| Comparison | Tines 3B | Tines Stories (the original Tines) |
|---|---|---|
| Positioning | AI-native, company-wide workflow platform | Visual designer for security and IT operations |
| How you build | Natural language + code-first | No-code visual building |
| Intended users | Staff across all departments + IT and security administrators | Security and IT operations teams |
| Going forward | Under development as a new product | Continued development and support in parallel |
According to the official blog, layering an AI-native experience on top of a no-code platform designed around 2018-era assumptions was not viable, so they rebuilt from scratch. Existing customers can keep using Tines Stories and try Tines 3B at no additional cost.
Who It’s For
- IT and security leaders who want to advance company-wide AI adoption but are struggling to rein in the automations and API key handling that employees create on their own
- Information systems departments that want to gather scattered AI agents, internal apps, and automations from every department into one place for inventory and cost management
- Developers who want to operate and monitor internal workflows from coding agents such as Claude Code
- Existing users already on Tines Stories who want to evaluate the AI-native next-generation environment
Summary
Tines 3B is a business workflow platform that starts from the premise that anyone can build with AI, and bakes in isolated execution, credential secrecy, auditing, and cost visibility from the outset. It is less a personal productivity tool and more a management foundation for organizations where AI adoption has raced ahead of governance. Since the free Explore edition lets you try the building experience, organizations where internal AI use is already spreading would do well to start with a small task. The one thing to keep in mind at the start of your evaluation is that it is a different product from the original Tines (Tines Stories).