Undes is built for tasks where a fast AI answer is not enough: code review, architecture decisions, bug investigation, CI checks, and long-running engineering analysis.
For first tests, local experiments and individual evaluation.
For solo developers and consultants using Undes regularly.
For organizations evaluating Undes beyond individual Pro usage.
Undes works with your own AI provider keys. Community focuses on OpenAI, Anthropic and Google. Pro is the paid track for a broader provider matrix.
The Community package is publicly installable from npm. It is intended for initial evaluation, local experimentation and understanding the Undes workflow.
Pro uses a separate licensed package for professional workflows: history, inspection, exports, provider expansion and local-model support as those capabilities ship.
The Pro provider-expansion track covers OpenAI-compatible endpoints such as OpenRouter, NVIDIA NIM and generic gateways, plus local model servers like Ollama, LM Studio and llama.cpp.
Some Undes runs can take longer than a normal assistant response. That is intentional: the product is optimized for evidence, critique, consensus and reviewable output.
Paid subscriptions will be processed through Paddle. Billing, invoices, taxes, subscription management and refund flows may be handled through Paddle checkout and receipts.
Team and Enterprise usage is not packaged as a public self-service plan yet. Contact us if you want to discuss organizational requirements.
FAQ
No. Undes is BYOK-first. You provide your own AI provider API keys, and model usage is billed by your selected AI provider according to their terms. The Community package is public, while Pro uses a separate licensed package.
Community is the public package for first tests and local experimentation. Pro is a separate licensed package intended for regular professional usage, session-aware workflows, run history, engineering memory, richer inspection, expanded provider support and email support.
Community is intentionally limited to first-party cloud providers: OpenAI, Anthropic and Google. OpenAI-compatible aggregators, hosted NIM endpoints, local model servers and generic gateways belong to the Pro provider-expansion track. Exact availability depends on the licensed package version and model qualification status.
Licensed packages make it easier to separate public evaluation from commercial functionality, license activation and support scope.
Yes. Undes is designed for workflows where analysis may run in the background: scheduled jobs, code review checks or CI/CD verification. For organization-wide usage, contact us first so the scope can be discussed directly.
There is no public on-premise package or installation path yet. Organization-specific requirements can be discussed directly, but no deployment model is promised on this page.
Yes. Subscription cancellation stops future renewals. Refunds and billing questions are handled according to the Refund and Cancellation Policy and applicable checkout terms.
Start with Community if you want to test the concept. Choose Pro if you plan to use Undes regularly for real engineering tasks. Contact us if you are evaluating Undes for a team or organization and need to discuss scope.
The best first test is not a toy prompt. Use a small real bug, architectural concern, or pull request where the final answer must be checked before merge.
