Discovery

AI to you help find the best parts across ALL the options without cracking open datasheets

TL;DR

Zenode's Discovery is what electronic component search should be in the era of AI. Type what you need in plain English, and the Agent searches across millions of parts from thousands of distributors and manufacturers, reads relevant datasheets, and returns cited, grounded answers with direct links to parts and catalog views. It's fast, it's sourced, and it's built for engineers who are choosing parts (not just buying them).

Introduction

Every electrical engineer knows the (old) drill: you need a part, so you head to a distributor catalog, pick a category, drill into subcategories, start setting filters, then start clicking through the top results to read documents. Repeat until you've got 42 Chrome tabs and 12 PDFs open on your second monitor.

It....works? 😅
Zenode's Discovery is a conversational AI search that sits on top of Zenode's unified catalog. You describe what you need, and the AI figures out where to look, what filters to apply, and which datasheets to read. It returns a sourced response with inline citations, part cards, catalog links, and follow-up suggestions (so you don't have to figure out what to ask next).

It's not a chatbot. It's closer to having a junior engineer who's read every datasheet in the catalog sitting next to you, pulling parts and explaining tradeoffs while you design.

Ways to Search in Discovery

Not all searches start the same way. Discovery supports two of the three main approaches:

1. Direct Part Number
Enter a full part number (e.g., LM317T) and jump straight to that part’s page.

2. General AI Query
Describe what you need in plain English.

• “3.3V LDO regulator in TO-220”
• “3-axis accelerometer with I²C, up to 30g”

The AI interprets your intent, applies filters, and shows relevant results. If it’s not perfect the first time, reword the query and try again.

3. Nonsense (What we don’t actually do 😅)
We’re built for electronic components, not consumer goods, GPUs, or economic predictions. Examples of what we don’t handle:

• “Speakers for a Mazda RX-7”
• “NVIDIA H200s”
• “Graph electricity prices in Norway for the next 5 years”

If it's not a part that would show up on DigiKey, Mouser, etc, and it's not made by Texas Instruments, TDK, etc, Zenode is not the right tool 😵.

What You See in a Response

Guided Discovery responses aren't just text. The AI generates several types of inline cards depending on what it finds:

• Part Cards show the image, MPN, manufacturer, key specs, and pricing for specific components. Click through to the full part page for datasheets and detailed specs.
• Catalog Cards show a category with the number of matching parts and active filters. Click "See parts in catalog" to jump into a filtered catalog view where you can sort, compare, and refine with parametric filters.
• Deep Dive Cards appear when the AI runs a deeper analysis across multiple parts' documentation. These are more expensive (token-wise) but surface insights you'd normally only get after reading multiple datasheets yourself.
• Sources appear at the bottom of every response as collapsible cards with clickable links. Inline citations throughout the text let you trace any claim back to its source. We deliberately avoided the dense inline-badge style (too noisy for technical content). Instead, sources are clean, scannable, and one click away.

@ References

You can also reference specific parts in follow-up queries using the @ symbol. Type @ followed by at least 4 characters of a part number, and an autocomplete overlay shows matching parts with their manufacturer and category. This is modeled after the Cursor IDE experience (if you've used it, you'll feel right at home).

Conversations maintain context across follow-ups, so you don't have to re-explain your requirements each time.

Why Discovery Matters

Component search happens 60+ times per design. It's a high-frequency, high-leverage, high Pain-In-The-Butt point in the every design workflow. Every minute saved per search compounds across a project. Here's what Guided Discovery changes:

• One search, all sources. No more tab-switching across distributors.
• Natural language, not category trees. Describe the function, not the taxonomy.
• AI reads the datasheets so you don't have to. The real work isn't reading one datasheet (that's 5% of the problem). It's reading hundreds during selection. The AI handles that.
• Sourced and cited. Every claim links back to a datasheet or spec sheet. No hallucination hand-waving.
• 15 minutes compressed to 2 minutes. That's the typical before/after based on our testing for the average part (critical parts will still take more time to figure out what you actually want, of course)
• Filters that actually work. Numeric ranges, dynamic counts, histograms (not string-matching garbage).
• Free catalog search, always. Even without credits, you've got a better search than most distributors.

Discovery FAQs

Quick answers to common questions about AI-powered component search, credits, and getting the best results.

Do I need an account to use Discovery?
Anonymous users get a handful of credits to try it out, but if you create an account and you'll get 25x as many.

How is Discovery different from Deep Dive?
They're different steps in the same flow. Discovery is the initial step to surface the most likely parts, then Deep Dive starts reading the documentation to analyze them in detail.

What if I type something that isn't a component?
The classifier catches most of these and routes them appropriately. If it slips through, you'll get a response explaining that Zenode is for electronic components only.

Why did my search take longer than expected?
Two common reasons: the AI is reading datasheets in the deep agent step (you can watch the thinking indicator to see progress), or it's backing off over-specified filters. You can stop the search early if you want faster but shallower results.

Can I buy parts directly through Zenode?
No, Discovery is about finding and comparing. Purchasing happens through distributors (we link to them).

How many parts does Discovery search across?
40 million+ searchable parts, merged from dozens of distributors and thousands of manufacturers. More than any single source.

Does it understand engineering shorthand?
Usually it does, although everyone has different shorthand. LDO, uC, SOT-23, 3V3, RDS(on) all work. The AI is trained on component terminology.

Can I share a search?
Every search has its own URL. Copy and share it.

What's the difference between Discovery and just using DigiKey/Mouser/etc?
Discovery searches across all major distributors at once, deduplicates parts, reads datasheets to understand what parts actually do (not just what category they're in), and cites its sources. DigiKey and Mouser's parametric search is not AI powered, and results are limited to just their catalog, their specs, their taxonomy.

What happens when I run out of credits?
Catalog search still works (always free, no AI cost).

How do I use the @ references?
Type @ in the search bar followed by at least 4 characters of a part number. An autocomplete menu appears with matching parts. Select one to reference it in your query (useful for searches like 'find me something similar to @LM317T').

How do I get the best results?
Start with a natural language description of what you need, including 2 to 4 key specs. More than 4 specs tends to narrow results too aggressively. If results aren't right, simplify and try again.

What if the AI puts me in the wrong category?
Rephrase your query. Adding one key differentiating term (like 'comparator' vs. 'op amp') usually fixes it.

Can I combine AI search with parametric filters?
Yes. Start broad with a natural language query, then refine with parametric filters on the catalog page. This combo is usually the fastest path to the exact right part.

How do I know if the AI's answer is trustworthy?
Check the sources. Every claim in the response links to a datasheet or spec sheet. Click through to verify. If you want a thorough analysis before committing to a part, use Deep Dive.