Support channel
Customer support is currently available through Instagram: @kumolumoai.
Kumo Lumo
Transparent AI routing for serious workspace use.
Kumo Lumo is a multi-model AI workspace, not a black-box chatbot. You choose a model route, or Kumo Agent selects one by task type, context size, point budget and capability. The product is designed to show what is happening before work becomes expensive, without Kumo itself using private workspace content for model training or selling user content.
Private workspace
Chats are treated as account workspace records. They are not published as public content and are used by Kumo Lumo to continue the user's own AI work, not to train Kumo models.
No Kumo training or sale
Kumo Lumo does not use workspace content to train Kumo Lumo models and does not sell user workspace content. This commitment does not mean model requests stay inside Kumo.
Provider transparency
Requests are processed by the model route selected by the user or by Kumo Agent. Only the context needed for the current request is sent to Poe and that route. Provider retention schedules and privacy terms still apply.
Cost and point clarity
The composer shows point estimates before sending. Final usage can change when a long context, attachment, media generation, or provider response is larger than estimated.
Estimate before send
Usage shown in workspace
Failed jobs should release reserved points
Files and generated media
Text conversations can remain available in the account. Uploaded files, images and generated media are handled as expiring assets retained for 7 days, so older downloads may become unavailable after the retention window.
Account chat history
7-day media retention
Clear expiry expectations
What Kumo Lumo is
A unified AI workspace for text, files, images, voice, video planning and commercial delivery workflows across multiple model families.
What Kumo Lumo is not
It is not a hidden social network, public prompt directory, or silent data broker. The value is model routing, workflow guidance, point estimates and a consistent workspace.
Trust FAQ
How can I contact Kumo Lumo support?
Customer support is currently handled through Instagram at @kumolumoai.
Which data is sent to model providers?
Only the prompt, uploaded references and conversation context needed to complete the current request are sent to the selected or routed provider. Different model providers can have different availability, latency and capability.
Does Kumo Lumo train AI models on my workspace?
Kumo Lumo does not use workspace content to train Kumo Lumo models and does not sell confidential prompts, files or generated outputs. Poe states that official full-shield model routes are not used by model providers for training, while their retention periods and privacy terms still apply.
Does the 7-day file policy mean zero upstream retention?
No. Seven days describes Kumo Lumo's short-lived workspace asset storage. Poe and each selected model provider can have separate retention schedules. Avoid sending secrets or regulated identifiers that are not needed for the task.
Can I use Kumo Lumo for confidential business documents?
Kumo Lumo is designed as a private workspace with necessary-context routing. Users should still redact unnecessary secrets, regulated identifiers and data they do not have permission to process.
Why do estimates sometimes differ from final usage?
Estimates depend on prompt length, selected route, attachments, context and output length. Final usage is settled after the response or generation job completes.
What happens when a provider route is unavailable?
External model routes can change or temporarily fail. The workspace is designed to show route status, fallback options and point handling instead of hiding provider availability.