deskree.com

What It Is

Tetrix is an “AI system knowledge infrastructure” product from Deskree. It is not just a code completion tool; instead, it maps code repositories, service relationships, dependencies, cloud resources, Kubernetes configurations, databases, and engineering documentation into a unified knowledge graph, giving AI system-level context when answering questions. Deskree was founded in 2021, with public company information showing its headquarters in Toronto, Canada, and operations in Austin.

Core Capabilities

Tetrix focuses on a deterministic “pre-mapped” structure: functions, classes, and dependencies are structurally parsed during indexing to create a persistent knowledge graph, rather than being inferred on the fly for each query. Queries are processed through five paths at once: code search, symbol traversal, dependency trees, semantic matching, and blast-radius analysis. Typical questions include: “Which services would be affected if this API changes?”, “Where is this function called across repositories?”, “Is a deprecated interface still used in production?”, and “Which service change may have caused the increase in latency?” It also offers Deep Research, multi-hop reasoning, MCP Integration, and an API. The Enterprise edition supports deployment in the customer’s cloud and bring-your-own-model setups.

Pricing and Privacy

The public page lists a free Community Edition, suitable for individuals and small teams, with self-hosting and the ability to connect repositories, documents, and infrastructure. The Terms of Service also list a Free plan with 1M tokens/month; Premium at US$50/user/month, including 10M organization tokens, with additional tokens at US$5/1M; and Enterprise with sales contact required, including priority support and volume discounts. Payments are processed by Stripe. On privacy, users retain ownership of their code, databases, infrastructure configurations, and business data. Deskree states that it does not use customer data to train models. However, AI features may send queries and relevant context to OpenAI and Anthropic, and third parties may retain data for up to 30 days for abuse monitoring.

Pros, Cons, and Best Fit

Its strengths are coverage across both code and infrastructure, making it useful for architecture reviews, impact analysis, migration planning, root-cause analysis, and onboarding in large systems. Its deterministic knowledge graph is also better suited to engineering dependency questions than ordinary semantic search. The limitations are that it requires access to highly sensitive systems, and permission governance, data compliance, and deployment costs may be significant. The AI output terms also make clear that accuracy is not guaranteed, so engineer verification is still required. Tetrix is better suited to mid-sized and large R&D organizations, platform engineering teams, and architecture teams, rather than users who only need personal code completion.

Access from China

The main materials do not state whether Tetrix is available in mainland China, whether it offers a Chinese interface, Chinese-language support, or local payment methods. Because it involves third-party AI providers such as OpenAI and Anthropic, network access from mainland China may be uncertain. Before enterprise adoption, companies should test network connectivity, compliance boundaries, and the payment process. Comparable alternatives include Sourcegraph Cody, GitHub Copilot, Cursor, Continue, Codeium, Glean, and others.