The Co-Analyst is built for institutional equity research - providing the precision of a terminal with the adaptability of AI. Get the data you need — fast, precise, and verifiable.
I’m Kris, co-founder of Hudson Labs with @piesauce. After years in finance and AI research, we’re excited to introduce the Co-Analyst — our high-precision AI platform built for institutional investors, hedge funds, and asset managers
Why we built this
We built the Co-Analyst after seeing the limits of traditional terminals like Bloomberg and FactSet. They’re great for standardized data, but analyzing guidance trends, niche operating metrics, supply chain exposure, product-level results etc. still require painfully manual workflows.
Generalist AI tools haven’t solved this. They break down in the workflows that matter most to investors — multi-document and multi-period analysis, guidance extraction, and numeric accuracy — failing to meet the reliability standards of institutional-grade research.
So we built the Co-Analyst:
- 🎯 Terminal-grade precision + AI adaptability - 🗂 Any metric, any source, any format — filings, transcripts, presentations, press releases - 📈 Soft guidance captured — every hedge, every forward-looking statement - 📝 Verbatim call summaries — no spin, no paraphrase - ⚡ 2 hours instead of 2 weeks to get up to speed on a new name
All direct-from-source, with no hallucinations, no prompting gymnastics.
Feedback:
- Any big wins you’ve had using AI in investment research? - Any frustrations? - What do you want from an AI tool for investing?
@kbenna Congrats on the launch, Kris! Co-Analyst looks sharp. The real test in finance isn’t summaries, it’s whether numbers reconcile across filings and calls. How are you handling trust and auditability so investors can rely on it like a terminal, not just an assistant?
@kbenna@tonyabracadabra Hi Tony, the co-analyst extracts numbers directly from relevant sources and does not rely on model memory. In order to enhance auditability and trust, answers come with fine-grained citations, with the ability for a user to select any part of the output and get a citation for it. Our citations allow you to see which paragraph/table of a document a particular number in the output is derived/extracted from.
@tonyabracadabra Great question, Tony. We have dynamic, AI-driven citations for both #s and qualitative statements. Everything's citable direct-to-source
Report
Want to know whether the hallucination generated by its AI is strong? After all, financial markets change all the time. If it can balance accessibility and depth, I believe that Hudson Labs can be suitable for investors and academic researchers, useful for quick investment decisions, and rigorous enough for academic research.
@wayne_appgrowing We're really strong. Here's a comparison vs. Perplexity, you can see we're able to offer reliability and precision where other AI tools struggle. Our customers have high expectations when it comes to accuracy - https://www.hudson-labs.com/post...
@wayne_appgrowing Hallucinations are still a problem with AI, even with the latest batch of reasoning models. At Hudson Labs, we have done extensive research on addressing hallucinations, and the Co-Analyst relies on several proprietary mechanisms to avoid them.
Report
Incredible seeing the accuracy of the co-analyst. I've used several tools and have always faced problems with data accuracy and retrieval.
As someone who’s wrestled with multi metrics and formats and different period analysis, this resonates. Bloomberg is great, but it’s not built for that kind of nuance. This looks like a big step forward. Hugeee Congrats!
@christyfea Thank you, Christy. Glad to hear this resonates
Report
Congrats on the launch! Really like how you’re tackling the “last mile” of research, extracting the metrics and nuance that usually fall through the cracks in terminals.
From my own experience, the real pain is never just getting numbers; it’s structuring them in a way that supports confident decisions. We’ve seen a similar thing on the strategy side with Escape Velocity AI: founders don’t just want data, they want frameworks that help them test assumptions and see if the story holds up.
Do you see early adopters leaning on Co-Analyst more for speed (cutting 2 weeks to 2 hours), or for surfacing insights they couldn’t realistically get before?
Making sure that the data is pulled directly from the sources seems obvious and somehow all the other products in the space, Perplexity included, don't do this.
@anson_kao Thanks Anson! We public equities only fwiw
Report
The auditability features are 🔥, direct-source citations for every number is exactly what’s been missing in finance-focused AI. This actually feels terminal-grade.
Report
Finally a product built for finance people, by finance + AI people. Most tools force generic models into finance use cases and miss the mark. This one actually feels like a big step forward for analysts and portfolio managers.
Replies
The Hudson Labs Co-Analyst
Hey Product Hunt 👋,
I’m Kris, co-founder of Hudson Labs with @piesauce. After years in finance and AI research, we’re excited to introduce the Co-Analyst — our high-precision AI platform built for institutional investors, hedge funds, and asset managers
Why we built this
We built the Co-Analyst after seeing the limits of traditional terminals like Bloomberg and FactSet. They’re great for standardized data, but analyzing guidance trends, niche operating metrics, supply chain exposure, product-level results etc. still require painfully manual workflows.
Generalist AI tools haven’t solved this. They break down in the workflows that matter most to investors — multi-document and multi-period analysis, guidance extraction, and numeric accuracy — failing to meet the reliability standards of institutional-grade research.
So we built the Co-Analyst:
- 🎯 Terminal-grade precision + AI adaptability
- 🗂 Any metric, any source, any format — filings, transcripts, presentations, press releases
- 📈 Soft guidance captured — every hedge, every forward-looking statement
- 📝 Verbatim call summaries — no spin, no paraphrase
- ⚡ 2 hours instead of 2 weeks to get up to speed on a new name
All direct-from-source, with no hallucinations, no prompting gymnastics.
Feedback:
- Any big wins you’ve had using AI in investment research?
- Any frustrations?
- What do you want from an AI tool for investing?
ScaryStories Live
@kbenna Congrats on the launch, Kris! Co-Analyst looks sharp. The real test in finance isn’t summaries, it’s whether numbers reconcile across filings and calls. How are you handling trust and auditability so investors can rely on it like a terminal, not just an assistant?
The Hudson Labs Co-Analyst
@kbenna @tonyabracadabra Hi Tony, the co-analyst extracts numbers directly from relevant sources and does not rely on model memory. In order to enhance auditability and trust, answers come with fine-grained citations, with the ability for a user to select any part of the output and get a citation for it. Our citations allow you to see which paragraph/table of a document a particular number in the output is derived/extracted from.
The Hudson Labs Co-Analyst
@tonyabracadabra Great question, Tony. We have dynamic, AI-driven citations for both #s and qualitative statements. Everything's citable direct-to-source
Want to know whether the hallucination generated by its AI is strong? After all, financial markets change all the time. If it can balance accessibility and depth, I believe that Hudson Labs can be suitable for investors and academic researchers, useful for quick investment decisions, and rigorous enough for academic research.
The Hudson Labs Co-Analyst
@wayne_appgrowing We're really strong. Here's a comparison vs. Perplexity, you can see we're able to offer reliability and precision where other AI tools struggle. Our customers have high expectations when it comes to accuracy - https://www.hudson-labs.com/post...
The Hudson Labs Co-Analyst
@wayne_appgrowing Hallucinations are still a problem with AI, even with the latest batch of reasoning models. At Hudson Labs, we have done extensive research on addressing hallucinations, and the Co-Analyst relies on several proprietary mechanisms to avoid them.
Incredible seeing the accuracy of the co-analyst. I've used several tools and have always faced problems with data accuracy and retrieval.
Congratulations to the team.
The Hudson Labs Co-Analyst
@abdullah_al_hayali Thanks Abdullah. That means a lot coming from you
Slay School
Congrats on the launch. When can i use this to make a gabillion dollars?!
The Hudson Labs Co-Analyst
@alijiwani1 Immediately - www.hudson-labs.com/#book-a-demo
As someone who’s wrestled with multi metrics and formats and different period analysis, this resonates. Bloomberg is great, but it’s not built for that kind of nuance. This looks like a big step forward. Hugeee Congrats!
The Hudson Labs Co-Analyst
@christyfea Thank you, Christy. Glad to hear this resonates
Congrats on the launch! Really like how you’re tackling the “last mile” of research, extracting the metrics and nuance that usually fall through the cracks in terminals.
From my own experience, the real pain is never just getting numbers; it’s structuring them in a way that supports confident decisions. We’ve seen a similar thing on the strategy side with Escape Velocity AI: founders don’t just want data, they want frameworks that help them test assumptions and see if the story holds up.
Do you see early adopters leaning on Co-Analyst more for speed (cutting 2 weeks to 2 hours), or for surfacing insights they couldn’t realistically get before?
Serverless Stack
Congrats on the launch! This looks great.
Making sure that the data is pulled directly from the sources seems obvious and somehow all the other products in the space, Perplexity included, don't do this.
The Hudson Labs Co-Analyst
@jayair Couldn't have said it better myself
Face Animator
Congrats Kris and team!! Hmm... who do I know in PE...
The Hudson Labs Co-Analyst
@anson_kao Thanks Anson! We public equities only fwiw
The auditability features are 🔥, direct-source citations for every number is exactly what’s been missing in finance-focused AI. This actually feels terminal-grade.
Finally a product built for finance people, by finance + AI people. Most tools force generic models into finance use cases and miss the mark. This one actually feels like a big step forward for analysts and portfolio managers.