Hey everyone, I've been putting MeetGeek through its paces for the last three months as part of our platform team's push to improve meeting hygiene and knowledge retention. Overall, I love the auto-transcription and the integration with our Notion wiki. But, I've hit a wall that's becoming a deal-breaker: **the search functionality is, frankly, inadequate for any serious, ongoing use.**
It feels like a bolt-on feature, not a core component. When I'm trying to find that specific conversation about the API gateway timeout configuration from a sprint planning session two weeks ago, I need precision. What I get is a fuzzy, keyword-matching mess that often surfaces entirely irrelevant transcripts.
Here’s a breakdown of my main pain points:
* **No Phrase or Proximity Search:** Searching for `"circuit breaker pattern"` doesn't respect the quotes. It returns every instance of "circuit," "breaker," and "pattern" scattered anywhere in the transcript, which is useless.
* **Poor Relevance Ranking:** The result ordering seems almost random. A passing mention of a keyword often ranks higher than a deep-dive discussion on the topic.
* **Zero Filtering by Participant:** After a large stakeholder meeting, I can't search for "budget" and filter to only see comments made by our engineering director. This context is critical.
* **No API for Search (that I can find):** This is a big one for my GitOps brain. I can't pipe meeting data into our observability stack or link findings from post-mortems to the exact meeting moment. The lack of a queryable API or webhook for search events locks the data in.
I've tried to work around it by meticulously titling meetings and adding agenda tags, but that's just putting a band-aid on a broken process. It defeats the purpose of a tool that's supposed to *reduce* manual overhead.
So, I'm actively looking for alternatives now. My core requirements are:
* Robust, Google-like search within transcripts (boolean operators, phrase search).
* Speaker-segmented search filtering.
* A strong API for automation and integration into our developer portal.
* Kubernetes-friendly (containerized deployment or a very secure SaaS model with fine-grained access controls).
Has anyone else run into this and jumped ship? I'm currently evaluating a couple of options:
1. **DIY with Whisper + Qdrant:** Using OpenAI's Whisper for transcription and a vector DB for semantic search. It's more work, but the control is tempting.
2. **Other SaaS tools:** Looking at platforms like Otter.ai, Grain, and Fireflies.ai, but their DevOps/integration story is unclear.
Any insights or shared experiences would be hugely appreciated. The dream is a meeting tool that feels like a first-class citizen in our Platform Engineering ecosystem, not a silo.
bw
Automate all the things.