Backlog
9Under consideration
Refund of the Starter subscription
Hi, Following up on my earlier message regarding attachment issues on efanovvasiliy@mail.ru. I want to clarify my position now that I've done further testing. I've since found that test files do upload successfully on both efanovvasiliy@mail.ru and efanovvasiliy@gmail.com, but only in a brand-new chat — the issue appears to be tied to specific chat sessions (possibly long ones) rather than the account itself. I also noticed that memory doesn't sync properly between the mobile app and the web client, even when logged into the same account on both. Given these findings, I don't think the Starter subscription purchase was the right call at this stage — I upgraded before I'd had a real chance to evaluate whether the product fits my workflow, and my first day was spent almost entirely on troubleshooting rather than actual use. I'd like to request a full refund of the Starter subscription charged on efanovvasiliy@mail.ru on 2026-09-15. I want to be clear that this isn't a complaint about Anuma as a product — I was recommended it by friends as reliable and well-built, and I'd like the chance to properly test the free tier, resolve the issues above, and make an informed decision about upgrading again once I understand the product better. Please let me know if you need any additional information to process this refund. Best regards, Vasily Efanov Please send your response to both efanovvasiliy@mail.ru and efanovvasiliy@gmail.com.
Limited Practical Capabilities of Multi-Agents
Hello Anuma AI Team, As an active user of Anuma AI, I am sharing a few observations and constructive suggestions to help improve the platform and enhance the overall user experience. 1. Slower Response Times When Deep Search is Disabled Issue: Ironically, the response time is noticeably slower when 'Deep Search' is turned off compared to when it is enabled. Testing: I tested this about 5 different times, and in almost every instance, disabling the Deep Search feature resulted in a delayed response. Suggestion: Typically, skipping the deep search process should yield faster results. Please review and optimize the internal routing/logic for queries when Deep Search is toggled off. 2. Limited Practical Capabilities of Multi-Agents (Feature Discrepancy) Issue: Having access to a wide variety of AI agents is a massive advantage for Anuma AI. However, the downside is that these agents cannot perform the core tasks they are originally capable of on their native platforms. Examples: While Gemini natively excels at generating AI Audio Overviews and advanced research tasks, selecting Gemini within Anuma AI does not allow for audio creation. Native models like GPT and Gemini can easily generate and output PDF files, but within Anuma AI, they either fail to do so or state that the feature is currently unavailable. Suggestion: Increasing the number of available agents is great, but it is equally important to improve integration depth. Please consider expanding the ecosystem so users can fully utilize each agent’s unique multimodal and file-generation capabilities. 3. Boosting User Acquisition via Blockchain Integration (ZetaChain & Fee Rebates) Proposal: To attract a larger user base and drive higher retention, I suggest deeply integrating blockchain mechanics, specifically by further activating ZetaChain. The Idea: A certain percentage (%) of the transaction/gas fees generated on the platform could be retained and converted into credits or mileage for the user. Users can then spend these accumulated credits directly within Anuma AI to pay for premium features, extra token usage, or specialized agent access. Expected Impact: This would heavily stimulate the ZetaChain ecosystem while providing a strong financial incentive ("the more you use, the more you earn") for users to stay locked into the Anuma AI platform. Thank you for your continuous hard work on this platform, and I look forward to seeing it evolve!
Public API for developers
A REST API so developers can build on top of Anuma — automate workflows, create custom integrations, and extend the platform.
Attachments not working on efanovvasiliy@mail.ru — requesting refund/credit transfer to efanovvasiliy@gmail.com
Description On the account efanovvasiliy@mail.ru, the Anuma interface shows a file successfully attached and sent, but the assistant never receives the file's actual content — only the user's text and, in some cases, screenshots of the interface itself. This happens consistently and does not depend on file format, size, or subscription tier. Steps to reproduce On September 15, 2026, I made multiple attempts to attach files in a chat under efanovvasiliy@mail.ru, testing the following formats: ZIP archive (containing two.json files),.doc,.docx,.txt, and PDF, ranging from 5 KB to roughly 10 MB. In every case the interface confirmed the attachment, but the assistant's context contained no file content at all — only text messages and screenshots I had manually attached. This behavior was fully reproducible and independent of chat length or prior messages. Hypotheses tested Free plan limitation — partially disproved: upgrading to Starter did not fix the issue; behavior remained identical. Unsupported file format — disproved: the exact same files, in the exact same Anuma interface, work correctly under a second account, efanovvasiliy@gmail.com (Free tier), where the assistant receives the content without issue. Tier-based size or type restriction — disproved: the gmail account on Free successfully opens files that the mail.ru account fails to open even on Starter. Root observation Two accounts belonging to the same owner — efanovvasiliy@mail.ru and efanovvasiliy@gmail.com — show different behavior for the exact same attachment pipeline. The failure appears to be account-specific, not a general Anuma limitation. Without backend access I cannot verify the exact cause, but the most likely explanation is a discrepancy in per-account feature flags or rollout configuration. Affected accounts and active subscriptions efanovvasiliy@mail.ru — Starter purchased on 2026-09-15, next billing cycle October 15, 2026. Attachments do not work. efanovvasiliy@gmail.com — Free tier. Attachments work correctly. Request I was recommended Anuma by friends as a reliable, modern, and very convenient service. In practice, my very first day of use has been one attachment failure after another, and my core use case — importing and analyzing JSON/PDF/doc archives — is currently unusable on the account where I actually paid. I'm requesting either: A full refund of the Starter subscription charged on efanovvasiliy@mail.ru, with the funds or equivalent credits applied to efanovvasiliy@gmail.com instead, or Direct transfer of the active Starter subscription and unused credits from efanovvasiliy@mail.ru to efanovvasiliy@gmail.com. Given that this issue effectively wasted my first day with your product — despite the high expectations I had going in — I would be willing to overlook the inconvenience if you could also offer a promo code or credit as compensation for the trouble. Attachments available on request: screenshots of interface attachments on mail.ru, screenshots showing the assistant's context lacking file content, and screenshots of the same files working correctly on the gmail account. Please send your response to both efanovvasiliy@mail.ru and efanovvasiliy@gmail.com. Best regards, Vasily Efanov
Testing Featurebase
Did this send
Bug with mobile app freezing randomly
Just use the mobile app for any extended period of time, it will eventually freeze when you work with enough chats and click enough buttons to the point I have to force close the app and restart again. Mobile app needs a lot of polish.
Bug with Council Mode between web and mobile
Hi team, Council mode does not sync properly with web and mobile. If a council mode chat was created in web first, it presents as an empty chat in mobile. If council mode chat was started in mobile first, it presents as some encrypted nonsense (see screenshot).
Bug with backing/syncing up between web and mobile
Hi team, I am using around 2.6 GB of storage, which isn’t that much and yet it’s oftentimes telling me that syncing has failed and so forth. It takes multiple tries to get this correct.
File Context is Not Shared Sequentially Across All Models in Multi-Agent Mode Body: I’ve noticed a critical issue with context/file sharing when running multiple agents simultaneously. When I uploaded a PDF manuscript (ASCOM-D-26-00177.pdf) and asked for an analysis, only a few models (like ChatGPT 5.5 and Gemini 3.1 Pro) successfully recognized and summarized the document. However, at the exact same time, other models (like Claude Sonnet 4.6 and Claude Opus 4.7) completely failed to read the uploaded file. Instead of analyzing it, they just outputted their default greeting templates (e.g., "What would you like me to analyze? Please paste your text..."). The core strength of the Multi-Agent mode should be "cross-validation"—allowing users to compare how different models analyze the exact same data to spot overlapping issues. Currently, because the file context isn't being properly fed to all selected models simultaneously, this collective intelligence feature is broken. Please review the multi-age
Hello Anuma AI Team, I am a frequent user of your service, and I would like to report a few bugs I encountered recently, along with a suggestion for a better mobile user experience. 1. File Upload Memory & Recognition Error (Infinite Loop Issue) Issue: After uploading a file and receiving the initial analysis, the AI often fails to remember the file in subsequent prompts. It either asks me to upload the file again or acknowledges that a file is there but says it cannot read it. Impact: Re-uploading the file within the same chat does not fix the problem. This forces me into an infinite loop where I have to open a completely new chat session and start over every single time. Please look into the context retention and file re-recognition logic. 2. 4-Agent Mode Bugs Upload Restriction Bug: If I select a specific agent first, the file upload feature becomes disabled. Strangely, if I upload the file first and then select the exact same agent, it works fine. Context Reset Bug: In this 4-agent mode, right after receiving the initial analysis result, asking any follow-up question completely resets the conversation history and context. 3. Mobile UI Improvement Suggestion (AI Tool Pros & Cons) Current Problem: On the PC version, it is very easy to see the pros and cons of each AI tool. However, this information is completely missing on mobile, making it difficult to choose the right AI on the go. Suggestion: It would be highly beneficial to add a small info button (like an 'i' icon) next to each AI's name on mobile. Tapping this button could reveal the pros, cons, and characteristics of that tool, matching the PC experience. I am sharing this in hopes of helping Anuma AI become an even better platform. Thank you for your time and effort in improving the service, and I look forward to seeing these updates! Best regards, [Minjun Kim]
Next up
Committed and queued
No Next up issues
In Progress
Actively being built
No In Progress issues
Done
13Recently shipped
Add short descriptions for each AI model
Anuma offers a wide range of advanced AI models, but first-time and casual users may not know which model is best for their specific task. It would be helpful to add a tooltip, info icon, or short description for each model explaining its strengths and best use cases, such as coding, creative writing, logical reasoning, speed, research, or general chat. Why it matters: Clear model guidance would help users choose the right tool faster, get better results, and feel more confident using Anuma.
Make Auto mode route prompts to the best model
Auto mode currently seems to default mostly to ChatGPT. It would be a major UX improvement if Auto could analyze the user’s prompt or intent and dynamically route the query to the best model for that task. For example, different prompts could be routed based on whether the user needs coding help, creative writing, deep reasoning, speed, research, or general assistance. Why it matters: Smarter Auto routing would reduce decision fatigue, improve answer quality, and make Anuma feel more intelligent and user-friendly.
Gpt image 2 and gpt image 2-thinking
Add gpt image 2 and gpt image 2 thinking.
Thinking and effort modes toggle.
A feature that lets you choose how hard a model things or the effort put in, however the more effort, the larger the credit deduction based on the model. For example say the normal thinking mode/effort prior to this update was medium: would deduct the same amount credits as the model does currently, however if the user selects low effort/low thinking it would deduct less, if the user chooses higher effort or thinking it would cost more points then per usual
Models like Claude Opus 4.8 cutting off mid response.
I utlized or attempted to in this case, to use the claude opus 4.8 model in help to aid me writing a python script for the sorce filmmaker program, I was heavily dissapointed after discovering the fact Claude Opus 4.8, mabye because of the new update to the Anuma website has developed a tendancy of halting its response mid way unlike on earlier versions of the website, resulting in a deduction of points and a half relevant answer.
Bug report and stablecoin payment request
Hi everyone, First of all, thank you for building such a solid product. I’m interested in subscribing, but I’m unable to do so through the usual channels. Are there any plans to add stablecoin payment options for the Starter/Pro plans? That would be a natural fit given Anuma’s crypto‑native foundation. In addition, I’m encountering two separate usability issues: Session‑initialisation hangs I often have to restart my browser and re‑enter my authentication key several times before the chat interface loads and my synced data appears. This seems tied to the “securing session” step, which hangs on load. I use a VPN as part of my standard privacy practice, and disabling it is not a viable workaround for me. PDF synthesis fails I attempted to generate a PDF file synthesis, but the output never appeared even though my credits were deducted. I lost 18 credits on a PDF synthesis that completely failed to render. This occurred on the mobile app; I have not tried the same operation in the browser to avoid further credit loss. Could you please look into these issues and let me know if there are upcoming fixes or work‑arounds? Thank you!
Native mobile app (iOS & Android)
A dedicated mobile app for Anuma with push notifications, offline access, and a native experience.
Housing & Leasing Agent
An AI agent dedicated to reviewing leases for red flags, drafting demand letters, analyzing rent increases against local market data, and guide you through security deposit recovery and move-out compliance.
App Builder — Create custom apps with AI
Build and deploy custom applications powered by AI directly within Anuma.
Can’t generate images but happy to continue taking my credits
☹️