AICHE +

R

Roam Research Integration

The short answer: open any Roam page or your Daily Notes, click into a block, press ⌃+⌥+R (Mac) or Ctrl+Alt+R (Windows/Linux), speak your thought for 30-90 seconds, and AICHE inserts clean text at your cursor. You add [[page references]] and block links afterward. The capture happens at speaking speed; the linking happens at editing speed.

The Capture-vs-Connection Tradeoff

Roam Research is built around one idea: every note should be a node in a graph. Write a thought, link it to related concepts with [[double brackets]], and over time a web of knowledge emerges that surfaces connections you never planned. The tool works beautifully for this.

The bottleneck isn't the tool. It's the input speed. Every block in Roam needs to be typed, and typing competes directly with thinking. You're reading a paper, and an insight hits you. You switch to Roam, click into a block, and start typing. Halfway through the sentence, you pause to remember the exact wording of the idea. It's already fading. You type a weaker version of the original thought, add a couple of brackets, and go back to reading. That 15-second interruption cost you both the idea's sharpness and your reading focus.

This happens dozens of times per research session. Each capture event requires a context switch from thinking mode to typing mode, and the tax accumulates. By the end of an hour of reading, you have 8 blocks that should be 20, each containing less detail than what was in your head when you started typing.

How Voice Capture Works in Roam

1. Open Roam Research in your browser.
2. Navigate to your Daily Notes page, or open any specific page where you want to add content.
3. Click into the block where you want text to appear. If you're starting fresh, just click below the last block.
4. Press ⌃+⌥+R (Mac) or Ctrl+Alt+R (Windows/Linux) to start AICHE recording.
5. Speak your thought fully. Don't stop to organize. Just say what's in your head - the insight, why it matters, what it connects to, what questions it raises.
6. Press the hotkey again. AICHE transcribes your speech, removes filler words, and inserts the text into the block.
7. After the text appears, go back through it and wrap relevant terms in [[double brackets]] to create page references. Add any block references or tags you want.

The two-pass approach matters. Pass one is pure capture at speaking speed. Pass two is linking and structuring at your own pace. Separating these tasks means neither suffers from trying to do both at once.

Workflows for Roam Users

Daily Notes as a Thinking Journal

Most Roam users start their day on the Daily Notes page. It's the default landing zone, automatically dated, and designed for unstructured capture. This makes it the ideal target for voice input.

Open today's Daily Notes. Press your hotkey. Speak for 2-3 minutes about whatever is on your mind. Projects you need to move forward, ideas that surfaced overnight, problems you're stuck on, articles you want to revisit. Don't filter, don't organize. Let AICHE capture the stream.

Afterward, break the transcribed text into individual blocks. Each block gets its own idea. Add [[page references]] where you spot connections to existing pages. These morning brain dumps become surprisingly valuable during weekly review, because you capture thoughts that normally evaporate before you sit down to type.

Literature Notes While Reading

When you're reading a book, paper, or long article, keep Roam open beside your reading material. Each time you hit a passage that sparks a reaction, switch to Roam, press your hotkey, and talk through it.

Example: "The author argues that spaced repetition only works when retrieval is effortful. Easy retrieval doesn't strengthen memory. This contradicts what I assumed about flashcard frequency. If the interval is too short, the card is too easy and the review is wasted. Connects to the desirable difficulties framework. My Anki settings might be too aggressive with short intervals."

That took 20 seconds to speak. Typing it would take 90 seconds, during which you'd lose your place in the reading and potentially forget the nuance of your reaction. After dictating, you add [[spaced repetition]], [[desirable difficulties]], and [[Anki]] as page references. The block is now a node in your graph.

Fleeting Thought Capture

Roam's graph improves with volume. The more atomic thoughts you capture, the more connections the graph reveals over time. But capturing a fleeting thought by typing requires enough motivation to open Roam, find the right spot, and type 2-3 sentences. That friction kills most fleeting thoughts before they reach the page.

Voice removes the friction. Open Roam (or switch to the tab that's already open), press your hotkey, speak for 10-15 seconds, done. The thought lands in your Daily Notes. You don't need to decide where it goes or what it connects to. The graph figures that out later when you add links during review.

Over a week, this habit produces 30-50 extra blocks that wouldn't exist if typing were the only input method. Some are noise. Some become the seed of your next project or the missing link between two ideas that have been sitting in your graph for months.

Why Voice Matches Roam's Design

Roam's philosophy is bottom-up. You don't plan a structure and fill it in. You capture atomic thoughts and let structure emerge through connections. This is the opposite of most document tools, which expect you to outline first.

Voice input is also bottom-up. When you speak, thoughts come in the order they occur to you, not in a pre-planned sequence. You mention connections as they surface naturally. The transcript is raw material, not a finished product. It needs editing, linking, and restructuring - which is exactly what Roam is designed for.

Typing encourages premature organization. You write a sentence, immediately wonder if it belongs in this block or a different one, restructure before you've finished capturing. Voice skips that. You get the content first, then shape it.

Heads-up: AICHE doesn't insert Roam-specific syntax like [[brackets]] or (()) block references. Speak your content in plain language, then add Roam markup during your editing pass. Trying to say "open bracket open bracket" during dictation breaks your flow and doesn't produce the right characters.

Pro-tip: Enable Content Organization in AICHE settings for literature notes. When you speak for 60-90 seconds about a complex passage, Content Organization structures the transcription into coherent points. This gives you cleaner raw material to break into blocks.

Result: Literature notes that took 20 minutes to type during a reading session now take 5 minutes to dictate. Daily Notes capture goes from 8 sparse blocks to 20 detailed ones. Your graph grows faster, surfaces more connections, and becomes the thinking partner Roam was designed to be.

Do this now: Open your Daily Notes in Roam, press your hotkey, and spend 2 minutes speaking about everything on your mind right now. Don't organize. Just capture. Then go back and add [[links]] to what you said.