Novexa (SMMY)
User: owners of small food & beverage and retail businesses in Baku (independent cafés, bakeries, salons, boutiques) who must post 3–5 times a week on Instagram and TikTok but have no marketing staff. Problem today: they either pay a freelance SMM specialist, which is a recurring monthly cost with days of turnaround and off-brand results, or post rushed phone photos themselves. Generic AI tools don't solve it: they write in English, invent prices and product claims, and output images that aren't the café's real product or setting. Outcome with Smmy: the owner fills in a brand profile once (tone, colours, audience, logo, product photos, verified prices and facts). Smmy then produces a batch of 3–5 ready-to-post items (image posts, carousels, stories, short videos) in Azerbaijani, English or Russian, with captions, hashtags and a posting schedule in Baku time. The owner reviews, edits by typing a plain instruction ("replace the latte with a glass of red wine, evening light"), approves, and downloads a package to post. A week of content goes from days of back-and-forth to one review session.
Preliminary score
- Value for the user19.5 / 25
A specific problem and a clear outcome from solving it.
- Prototype and use of AI21.5 / 30
A working core scenario, and what AI actually contributes to the result.
- Quality testing11.5 / 20
Tests, examples of failures, and comparison with the current approach.
- Feasibility9.0 / 15
Data requirements, running costs and a clear next step.
- Originality6.0 / 10
A meaningful difference from common solutions.
What the judges wrote
- GPT-6 AstraPreliminary71/100
- Strengths
- Your brand profiles, reference-photo editing, and Azerbaijani/English/Russian output fit a specific need among small Baku businesses without marketing staff. Your planner-to-image-to-video workflow gives AI a substantive role, while deterministic layout edits avoid unnecessary generation costs. Your reported regression tests address practical risks, including duplicate paid jobs, stale results, user isolation, and approval of unfinished videos.
- To improve
- Your live-AI testing section contains example wording rather than measured results; replace it with actual outputs, instruction-following rates, latency, and checks for invented claims. Your adoption case would benefit from an owner pilot comparing time, editing effort, and cost against the current posting workflow. Your feasibility plan should quantify per-batch generation costs, including retries, and specify spending limits and the next pilot milestone.
- Claude Opus 5.5Preliminary64/100
- Strengths
- Your user and pain are concrete: Baku café and retail owners who must post several times a week in Azerbaijani, Russian or English without marketing staff. Your pipeline gives AI a clear, essential job. A planner turns free-text, negated instructions into a structured edit plan, the model edits the real reference photo instead of inventing an unrelated image, and an image-to-video step animates the edited frame. Your engineering tests are unusually thoughtful for a hackathon. They prevent paid jobs from being submitted twice, keep stale AI results from overwriting newer edits, isolate users from each other and fail clearly when the API key is missing. You also reported real failures, such as the fake zoom-'video' and the 401 deployment bug, and how you fixed them.
- To improve
- Your 'live AI test results' section still holds template placeholders ('3 of 5 runs…', 'X minutes'), so there is no evidence yet of how often edits follow the instruction or how long generation takes. Run a small, systematic set of real edits and report success rates, failure examples and timings. Put numbers on your adoption metric: compare cost per week of content and turnaround against a Baku freelance SMM specialist. Add the per-post API cost for image editing and the 1080p video. Name a concrete next step, such as a pilot with a few real cafés, and add a deck or a public repository so judges can verify the build.
The team
Yavar JafarliLead
data
Eltun Tagiyev
designer
Huseyn Isayev
business