MakinaRocks
Also known as: MakinaRocks Runway, Runway AI OS, DrawX, makinarocks.ai
Industrial AI company selling two halves: Runway, an AI operating system that handles identity, permissions, GPU allocation and audit logging once at the platform level so teams can deploy pre-installed, catalog or bring-your-own AI applications on top, built to install in air-gapped networks; and Specialized AI, four agent lines for expert augmentation, control, predictive operations and machine vision, plus DrawX for engineering drawings.
MakinaRocks builds AI for industrial operations, and sells it in two halves.
Runway is its platform, positioned as an AI operating system, the layer that sits between GPU servers, storage and networking and the AI applications an enterprise wants to run on them. The argument is that most enterprise AI trouble is operational, not technical. Every new application has its own authentication, its own access rules and its own infrastructure setup, and across dozens of teams that becomes unmanageable. Runway handles authentication, access control, resource management and security once, at the platform level, and lets teams deploy what they need on top.
Everything that runs on Runway is treated as an app. A set arrives pre-installed and works from the moment the platform is on: source control, experiment tracking, pipeline orchestration, GitOps deployment, secrets management, LLM tracing, a playground for managing language models, and a chat interface, all behind a single sign-on so there is no separate login per tool.
An app catalog offers more to install on demand, such as development environments, vector databases and a visual builder for retrieval pipelines and multi-agent workflows. Anything else can be deployed as a Helm chart, including a company's own model-serving API or internal tools, and whatever route an app arrives by, the same authentication, permissions and resource governance apply to it automatically. How the pieces connect is left to the team instead of fixed by a pipeline built into the platform.
Two things shape who buys it. The first is governance: single sign-on, role-based access control, fine-grained permissions with built-in or custom roles, and centralized audit logging that captures every operation, all administered from one control plane and covering every application on the platform. The second is where it will run.
Runway is built to install in air-gapped and disconnected networks with no external dependencies, using automated offline packages and private package mirrors, and the full set of pre-installed tooling works there too. The company reports cutting air-gapped deployment from months to weeks. Alongside that it manages GPU supply, using fractional slicing and reclamation of idle capacity to raise utilization.
The second half of the business is Specialized AI: agent lines for expert augmentation, control, predictive operations and machine vision, aimed at automotive, defense, heavy industry and manufacturing, plus DrawX, a separate product that turns engineering drawings into decisions. Customers named on the company's own site include defense and public-sector institutions, insurance and credit bodies, and manufacturing groups.
MakinaRocks is headquartered in South Korea, publishes in Korean, English and Japanese, and maintains an investor-relations section. Runway is at version 2.0. There is no public pricing. The product is sold through demo requests and sales contact.
Vendor details
Canonical URL
https://www.makinarocks.ai
Category
Enterprise operations agent
Company status
independent
Use cases & customers
In practice
An industrial robot is heading toward failure and your team finds out when it stops. MakinaRocks builds predictive-maintenance models on Runway that catch the anomaly while the line is still running.
Your plant runs on an air-gapped network, so cloud AI tools are off the table. Runway installs in disconnected networks from offline packages and private mirrors, and its full set of pre-installed tools works there too.
Dozens of teams each stand up their own AI tools with their own logins and access rules. Runway puts every app behind one sign-on, applies role-based permissions to every application on the platform, and logs every operation centrally.
Sources & related URLs
Related / legacy domains
Agentic Index coverage score
9.0 / 14 capabilities · 64%
| Integrations & Tool Calling | Partial |
|---|---|
|
Real time industrial data flows into the platform, but nothing writes into an outside system. The PLC connection is the real support, and it is a read. The other named counterparties, GPU servers, storage and network, are the platform's own substrate. The manufacturing pitch to integrate AI across design, production, quality and operations into a single framework is about consolidating the customer's AI systems, not about agents reaching outside systems. The internal integrations are real. The platform brings Gitea, MLflow, Airflow, Argo CD, OpenBao, LLM Playground, Chat and Langfuse under one identity plane, and accepts any workload via Helm chart. That is integration into the platform, not out of it. There is no authenticated action in an outside system, and no API agent that triggers external system actions, DB agent that queries databases in natural language or retrieval agent that consults manuals. Sourcemakinarocks.ai/en/product/runway/ industries section, makinarocks.ai/en/blog/what-is-ai-os/ platform apps section, PLC use-case page carried from the July passread 2026-09-14 |
|
| Workflow Orchestration | Full |
|
Two orchestration surfaces ship with the platform, one of them multi agent by name, and the sequence is left to the customer. Airflow for orchestration is a platform app present from installation, and the app catalog carries Langflow for building RAG pipelines and multi-agent workflows without writing orchestration code, a visual builder for multi agent work. The composition model is the product's central claim. Applications are not locked into a fixed pipeline, every component is independent, and the team decides how they connect, for example running experiments in JupyterLab, logging results to MLflow and deploying via Argo CD, or building a RAG pipeline in Langflow and serving it directly through Chat. Airflow and Langflow are open source components MakinaRocks packages, integrates under one identity plane and ships as part of the platform. MakinaRocks also describes an enterprise AI operating system that orchestrates specialized agents for anomaly detection, predictive maintenance and process control, but names no mechanism for it. There is no orchestration runtime of MakinaRocks' own and no workflow versioning, so this is shipped orchestration tooling, not a model of how agents coordinate. Sourcemakinarocks.ai/en/blog/what-is-ai-os/ platform apps, app catalog and own-the-os sectionsread 2026-09-14 |
|
| Knowledge Grounding & RAG | Partial |
|
The components to build a knowledge store ship with the platform, but none is operated over the customer's data. The app catalog offers vector databases including Milvus, Qdrant and Chroma DB and Langflow for building RAG pipelines, and the unified runtime lists vector databases among the connected components. These ship with the platform and are governed by it. The customer gets an open source database engine and pipeline builder, one tap away, inside the platform's identity plane. A vector database is the means to build a knowledge structure, not the structure itself. Here the index arrives empty. Nothing ingests the customer's manuals, drawings or maintenance histories, no source is indexed, refreshed or permissioned by the product, and there is no schema, refresh path or ownership of what is in the store. There is no citation visibility either, because MakinaRocks operates no retrieval. Shipping Milvus is shipping the empty shelf. Sourcemakinarocks.ai/en/blog/what-is-ai-os/ app catalog and unified runtime sectionsread 2026-09-14 |
|
| Human Oversight & Guardrails | Partial |
|
Runway controls who may run what, but no gate holds an agent's action for human approval. The platform controls who can access which apps and data from a single control plane, administrators can use built-in roles or define custom permission sets, and fine-grained permission controls extend to every AI application running on Runway. MakinaRocks compares this to phone app permissions, where settings control exactly what each app can access, such as camera, location, microphone and contacts, and the OS enforces those policies centrally. Permissions bound who may run what, not what must be approved before it commits. A role that prevents an engineer from deploying an app is an access decision taken once, at configuration time, and there is no approval queue, human in the loop step or hold for review between an agent's output and its effect. The line that the freedom to install any app does not mean anything goes is about which software may run, not which actions may proceed. No MakinaRocks page says agents suggest setpoint adjustments for human or system approval, and the defense pitch of absolute governance and security protocols is about control of AI deployment, not a person approving an agent's setpoint. Sourcemakinarocks.ai/en/blog/what-is-ai-os/ governance section, makinarocks.ai/en/product/runway/ industries sectionread 2026-09-14 |
|
| Security, Identity & Governance | Full |
|
ISO certification badges sit alongside a deep identity and permission surface, though the certificates themselves are not available. Every page carries ISO 27001 and ISO 9001 badges, a Korean Good Software certification mark and a defense innovation program designation. The controls go deeper. Security is enforced across all layers through Keycloak-based SSO and RBAC, with the platform handling authentication, access control, resource management and security once, at the platform level. Runway 2.0 handles this through SSO, RBAC, fine-grained permission controls and audit logging, all managed from one place, and administrators can use built-in roles or define custom permission sets for their organization. This governance layer extends beyond models and datasets to every AI application running on Runway, and the platform apps include OpenBao for secrets. The gap is verification. The ISO badges are images with no certificate number, certifying body, scope statement or issue or expiry date, and there is no trust center or security page, the only legal page being a privacy policy, so a customer cannot verify the scope of what was certified. Sourcemakinarocks.ai/en/product/runway/ governance section and site footer, makinarocks.ai/en/blog/what-is-ai-os/ governance sectionread 2026-09-14 |
|
| Observability & Auditability | Full |
|
Every platform operation lands in a centralized audit log, and an LLM tracing tool ships pre installed. Runway automatically captures every operation in centralized audit logs, which MakinaRocks says ensures 100% compliance, and its governance covers who can access which data, what actions have been taken, whether the entire operational trail is auditable and whether every action is traceable, all managed at the OS layer. Capturing every operation centrally, as a property of the platform and not a per-app afterthought, makes a record at run level. The second surface is Langfuse, one of the pre installed platform apps. Langfuse is an LLM tracing and observability tool, so the platform ships a surface on which an LLM application's individual runs can be inspected, which lifts this above a bare traceability claim. What is observed is the AI applications and agents running on Runway. The industrial monitoring the specialized agents perform on plant equipment is the product's output, not a record of the agents. There is no retention period, export path or SIEM integration, and the audit claim is stated at platform level, not shown as a log schema. Sourcemakinarocks.ai/en/product/runway/ governance section, makinarocks.ai/en/blog/what-is-ai-os/ platform apps and governance sectionsread 2026-09-14 |
|
| Memory & State Persistence | Not documented |
|
There is no agent memory of any kind. MakinaRocks describes the platform component by component, from shared GPU and storage services to the app catalog and audit logging, and no store of agent state appears in it, with no scope, lifetime or carry over between runs. Continual learning applied to its models is state accumulating in model weights, which is retraining, not something an agent reads back as context. The vector databases in the app catalog are knowledge stores over the customer's documents and data. A vector index is a retrieval structure, not a memory of what an agent did or decided, and offering three as installable options makes clear these are components the customer fills. Even conversation state is missing. The platform ships Chat as an AI interface, but there is no thread persistence, history retention or context carried between sessions, and storage as an infrastructure service is disk, not state. Sourcemakinarocks.ai/en/blog/what-is-ai-os/ read in full, makinarocks.ai/en/product/runway/ read in fullread 2026-09-14 |
|
| Deployment & Data Residency | Full |
|
Installation in air gapped and disconnected networks is a headline capability, described in working detail, alongside cloud and on premises options and edge execution on industrial computers. Runway offers instant deployment for air-gapped environments, deploying AI securely in restricted or disconnected networks, with automated import pipelines and private package mirrors for instant execution with zero external dependencies. The mechanism, automated offline packages and private mirror repositories, cuts deployment in air-gapped networks from months to two weeks by MakinaRocks' account. A vendor that has built package mirroring to make offline installation fast is operating in disconnected networks, not describing an aspiration. MakinaRocks treats it as a design requirement of the category, noting that in regulated industries like defense, manufacturing, public sector and financial services external network access is often restricted and AI applications need to run reliably there too. The capability extends to the tooling as well as the runtime. The full pre-installed suite works even in air-gapped environments, and the defense positioning is to deploy and control AI in air-gapped and high-security environments with zero external reliance. There is no region list or residency menu for a hosted option and no installation or sizing documentation, but the route into the customer's environment is described repeatedly. Sourcemakinarocks.ai/en/product/runway/ air-gapped feature, results and industries sections, makinarocks.ai/en/blog/what-is-ai-os/ air-gapped support sectionread 2026-09-14 |
|
| Prebuilt Agents, Templates & Packs | Partial |
|
The browsable app catalog holds infrastructure, not agents. MakinaRocks describes it in app store terms, with vetted applications ready to install with a tap. Six installable entries are shown (Chroma DB, Code Server, JupyterLab, Langflow, Milvus and Qdrant) plus eight pre installed platform apps, and every one is a development environment, a database or a piece of engineering tooling. A catalog of vector databases can be browsed, but it is not a set of agents a customer adopts. The units that would qualify sit elsewhere. Four agent lines appear under Agents (Expert Augmentation, Control, Predictive Operations and Machine Vision), each with its own durable URL, alongside a separate named product agent, DrawX, which turns drawings into decisions with an AI agent. Nothing settles whether those four are adoptable units or four descriptions of one capability set, so they are named machinery and named agent types, not a catalog of adoptable units. Sourcemakinarocks.ai/en/blog/what-is-ai-os/ app catalog and platform apps sections, makinarocks.ai primary navigation read across three pagesread 2026-09-14 |
|
| Triggers & Channel Coverage | Partial |
|
A scheduler and a chat interface ship with the platform, and agents are described acting on real time PLC data, but the platform has no event seam of its own. In one manufacturing use case, agents monitor production conditions continuously from real-time PLC data and act as conditions change. Airflow for orchestration is one of the eight platform apps available from installation, even in air-gapped environments, and Airflow is a scheduler, shipped and not just described. Chat is a third channel, also pre-installed, and MakinaRocks' illustration of composition ends with building a RAG pipeline in Langflow and serving it directly through Chat. A conversational surface is a channel, though a human-invoked one, so it widens coverage without evidencing autonomy. There is no webhook, event subscription, inbound queue or trigger configuration surface. What exists is a scheduler the customer wires up and a continuous data feed described in a single use case. Channel coverage is thin, with one real-time industrial feed, one scheduler and one chat interface, all within a single deployment. Sourcemakinarocks.ai/en/use-case/runway-manufacturing-ai-system-with-real-time-plc-data/ carried from the July pass, makinarocks.ai/en/blog/what-is-ai-os/ platform apps sectionread 2026-09-14 |
|
| Model Flexibility & Routing | Full |
|
The customer's administrators choose and serve the models, through a model management app that comes pre installed. LLM Playground for managing large language models is one of eight platform apps present from installation, inside MakinaRocks' own identity and lifecycle plane, where everything is integrated through Keycloak-based SSO with one Runway login for everything. A model management surface handed to the customer's administrators answers who chooses, and the playground is MakinaRocks' own app, not a tool from another company. The unified runtime connects model configuration, serving infrastructure, vector databases and orchestration tooling within a single platform, and a MakinaRocks product update covers permissions, LLM serving in a closed network and experiment setup. Serving language models inside a closed network is the customer running models of their choosing on their own hardware with no provider in the path, and the pre-installed suite works even in air-gapped environments. A customer can also deploy a proprietary model serving API through the Helm path, which is permitted, not shipped. This covers the Runway platform, and the Specialized AI agent lines may run on a model MakinaRocks fixes. MakinaRocks publishes no detailed model list, routing policy or selector interface. Sourcemakinarocks.ai/en/blog/what-is-ai-os/ platform apps and unified runtime sectionsread 2026-09-14 |
|
| APIs, SDKs & MCP Extensibility | Partial |
|
Customers can deploy their own software onto Runway, but nothing lets an outside system call the platform. The extension path is open and unbounded. Runway supports deploying any workload via Helm chart, whether an open-source tool not in the catalog, a proprietary model serving API, an internal analytics dashboard or a custom admin tool, with no vendor-curated boundary. Governance follows the extension instead of blocking it. Wherever an app comes from, the catalog, a Helm chart or joint development with MakinaRocks, the same authentication, access control and resource governance applies automatically. A named, standard packaging format with a deployment path and automatic policy inheritance is a real extensibility surface, not a marketing claim. The direction is inward, not outward. Helm charts let a customer put software onto the platform, but there is no API or SDK that lets an external system or another vendor's agent drive Runway. There is no REST API, CLI reference, MCP server, webhook or documentation host. A proprietary model serving API deployed by the customer is the customer's API running as a workload and says nothing about whether the platform itself is callable. An enterprise platform of this kind may well expose APIs under contract, but MakinaRocks asserts none. Sourcemakinarocks.ai/en/blog/what-is-ai-os/ custom deployment and governance sections, makinarocks.ai navigation and footer read across three pagesread 2026-09-14 |
|
| Testing, Debugging & Optimization | Full |
|
A pre deployment harness comes pre installed, with experiment tracking in MLflow ahead of a GitOps release. MLflow for experiment tracking is one of the platform apps present from installation, and MakinaRocks describes the working pattern as running experiments in JupyterLab, logging results to MLflow and deploying via Argo CD. Experiments logged and compared before a GitOps deployment put a change under test, and it is the sequence, not the tool, that makes it one. The release path sits alongside it, Argo CD for GitOps-based deployment with Gitea for source control, so a model or application change moves from a tracked experiment through a versioned repository into a deployment. MakinaRocks names experiment setup as a shipped capability in a product update, and the full suite works even in air-gapped environments, so the harness is not a cloud-only convenience. MLflow, Argo CD and Gitea are open source components MakinaRocks packages, governs under one identity plane and ships as the product. The object is what runs on the platform, which for a customer deploying MakinaRocks' own specialized agents includes those agents. The tagline of managing the full model lifecycle from planning and risk modeling through operations with continual learning describes lifecycle management, which is not a harness, and continual learning is absorbed into the model. There are no scored test cases, golden set, accuracy figure or published evaluation methodology of MakinaRocks' own. The harness is named, but its use is not measured. Sourcemakinarocks.ai/en/blog/what-is-ai-os/ platform apps and own-the-os sectionsread 2026-09-14 |
|
| Browser & Computer Use | Not documented |
|
The agents act on industrial systems and machine data, with no screen in the loop and no browser or computer use. The action path is data and infrastructure, such as real time PLC feeds, Helm deployed workloads and vector databases, with no rendered control to target. Machine Vision, one of the four agent lines, inspects physical parts and surfaces through cameras, which is seeing, not driving a computer interface. Breakage when UI elements change has no meaning against PLC tags, Helm charts and industrial cameras. JupyterLab and Code Server in the app catalog are development environments a person sits in front of, and Chat is an interface to the agents, not one the agents operate. MakinaRocks describes the architecture in enough detail that an undocumented screen-driving path is implausible. Sourcemakinarocks.ai/en/product/runway/ read in full, makinarocks.ai/en/blog/what-is-ai-os/ app catalog section, site navigationread 2026-09-14 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
Enterprise pricing. Contact sales.
Related vendors
- 11th Estate — Agentic platform whose AI engine scans markets, matches an…
- Abacum — FP&A platform for growing companies whose AI Modeler builds, edits…
- Adopt AI — AI native accounting and bookkeeping firm whose agents run…
- Aera Technology — Decision intelligence platform where an always on agent executes…
- Akro AI — On premise operational intelligence that automates document heavy…
- alfred_ — AI executive assistant that triages the inbox overnight and scores…
Alternatives to MakinaRocks
The closest documented capability profiles to MakinaRocks among enterprise operations agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Uniphore9.5 / 14Fuller documented coverage on Knowledge Grounding & RAG and Prebuilt Agents, Templates & Packs
- Scaled Cognition8.0 / 14Fuller documented coverage on Integrations & Tool Calling
- Box10.5 / 14Fuller documented coverage on Integrations & Tool Calling and Knowledge Grounding & RAG
- Forest10.5 / 14Fuller documented coverage on Integrations & Tool Calling and Human Oversight & Guardrails
- HireVue7.5 / 14Fuller documented coverage on Integrations & Tool Calling
- Atomicwork12.0 / 14Adds documented Memory & State Persistence
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded