Pressing enter key doesn't enter the prompt and instead goes to next line, You always have to press the arrow button, might be nitpick but I feel adding it would give better user experience for pc/laptop users

Hi! I'm pretty new to bun and cloudflare. But this application is amazing so I really want to run it on local.
I followed the setupguide.md and I was able to setup until step 7.
But for step 8, when executing the following command, I encountered an issue: "bun run". During the process, there was an error message:

"$ bun dev
$ dotenv -- turbo run dev --filter=!cf-ai-backend
node:events:491
throw er; // Unhandled 'error' event
^

Error: spawn turbo ENOENT
at ChildProcess._handle.onexit (node:internal/child_process:285:19)
at onErrorNT (node:internal/child_process:483:16)
at process.processTicksAndRejections (node:internal/process/task_queues:82:21)
Emitted 'error' event on ChildProcess instance at:
at ChildProcess._handle.onexit (node:internal/child_process:291:12)
at onErrorNT (node:internal/child_process:483:16)
at process.processTicksAndRejections (node:internal/process/task_queues:82:21) {
errno: -2,
code: 'ENOENT',
syscall: 'spawn turbo',
path: 'turbo',
spawnargs: [ 'run', 'dev', '--filter=!cf-ai-backend' ]
}

Node.js v19.0.0
error: script "dev" exited with code 1"

How can I solve this error? Also, I can share the information until step 7.

I added some memories, and then asked a follow-up question relevant to one of them.

Instead provides some vague response(tho it quotes the relevant memory).

Connect (& Arrange) all data with some (automated) AI based logic (could be dynamic &/ expandable).
Output may be in form of canvas/knowledge graph like in obsidian/something similar. (Data Arrangement Algorithm Goal: Any data should be reachable with minimum node clicks)
So whenever user has more time to look into, he/she can explore/scroll thru data from knowledge graph.

Currently, the only way to add notes is from here -

Upon recieving feedback about this, I found that the notes feature is actually very helpful and should be a button in the sidebar itself. also because the sidebar is pretty empty right now!

Can this just download your twitter bookmarks and save them locally? To then be integrated with another project I'm working on?

Ensure everything required to make this work is already on the chatWindow page. It involves making a simple fetch call.

README says to use https://chromewebstore.google.com/detail/supermemory/afpgkkipfdpeaflnpoaffkcankadgjfc?hl=en-GB&authuser=0

trying to sign into that extension takes you to https://supermemory.dhr.wtf/api/auth/signin which doesn't resolve

onboarding should have a tutorial / CTA on how to import existing data

<button
          onClick={() => {
            signIn("google");
          }}
          className="flex flex-row gap-4 items-center justify-center px-8 py-4 rounded-md text-black font-semibold font-raleway bg-white"
        >
          <Image src={"https://upload.wikimedia.org/wikipedia/commons/thumb/c/c1/Google_%22G%22_logo.svg/768px-Google_%22G%22_logo.svg.png"} alt="" width={30} height={30} />
          Sign in with Google
        </button>

this button should be ok ig I've just picked it from one of my projects, I've changed the image src to some random google logo I found online.

Will there be a Safari extension anytime?

Please allows Safari iOS users to use it... this is too cool..

Please port the extension to firefox

Hi!

Thanks for creating this awesome web application! I hope you can support Claude models in addition to OpenAI models. Claude models are great, especially the Opus model.

the extension and auth text available on sidebar do not provide any functionality

Please make a firefox version of the chrome extension

Hi,
I saw your project and I was attracted by it. You seem to find a common problem among people (especially developers) with an uncountable to read list. After reading the readme, I couldn't understand your solution to the problem actually, I reread the documentation but couldn't find the solution.

My question is, how do you want to overcome the problem? Just by storing it in a database and starting chatting with that?

Wanna know more about what you are thinking about and what the roadmap of the project is.

Thanks.

Provide a button for batch import browser bookmark URLs
Chrome browser(almost all mainstream browsers ) bookmarks are exported as an HTML file, so i think we can get the URLs by reading all the href attributes of <a> tag.

Unhandled Runtime Error
Error: Hydration failed because the initial UI does not match what was rendered on the server.
See more info here: https://nextjs.org/docs/messages/react-hydration-error

Did not expect server HTML to contain a in

Since it(SM) will have all data, while creating any new note/writing draft in note taking tool (like obsidian or anything else), it would be super useful if it can work as autocomplete/suggest related snippets from existing data.

Steps to reproduce

1. I click on login through the extension
2. Then I sign through the app
3. But still it says to login
4. Because of that whatever website I try to bookmark, nothing gets saved and my usage is shown zero

Happy to post any logs if required to solve this!

Thanks Dhravya, curious to check your app :)

Hi! I was able to run the application locally, but when I login with google account the error below occurred.

[auth][error] AdapterError: Read more at https://errors.authjs.dev#adaptererror
web:dev: [auth][cause]: TypeError: Cannot read properties of undefined (reading 'prepare')
web:dev:     at SQLiteD1Session.prepareQuery (webpack-internal:///(rsc)/../../node_modules/drizzle-orm/d1/session.js:29:30)
web:dev:     at SQLiteD1Session.prepareOneTimeQuery (webpack-internal:///(rsc)/../../node_modules/drizzle-orm/sqlite-core/session.js:76:17)
web:dev:     at SQLiteSelectBase._prepare (webpack-internal:///(rsc)/../../node_modules/drizzle-orm/sqlite-core/query-builders/select.js:586:88)
web:dev:     at SQLiteSelectBase.get (webpack-internal:///(rsc)/../../node_modules/drizzle-orm/sqlite-core/query-builders/select.js:604:17)
web:dev:     at getUserByAccount (webpack-internal:///(rsc)/../../node_modules/@auth/drizzle-adapter/lib/sqlite.js:124:18)
web:dev:     at acc.<computed> (webpack-internal:///(rsc)/../../node_modules/next-auth/node_modules/@auth/core/lib/init.js:178:30)
web:dev:     at Module.callback (webpack-internal:///(rsc)/../../node_modules/next-auth/node_modules/@auth/core/lib/actions/callback/index.js:53:39)
web:dev:     at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
web:dev:     at async AuthInternal (webpack-internal:///(rsc)/../../node_modules/next-auth/node_modules/@auth/core/lib/index.js:39:24)
web:dev:     at async Auth (webpack-internal:///(rsc)/../../node_modules/next-auth/node_modules/@auth/core/index.js:126:34)
web:dev: [auth][details]: {}
web:dev: [auth][error] AdapterError: Read more at https://errors.authjs.dev#adaptererror
web:dev: [auth][cause]: TypeError: Cannot read properties of undefined (reading 'prepare')
web:dev:     at SQLiteD1Session.prepareQuery (webpack-internal:///(rsc)/../../node_modules/drizzle-orm/d1/session.js:29:30)
web:dev:     at SQLiteD1Session.prepareOneTimeQuery (webpack-internal:///(rsc)/../../node_modules/drizzle-orm/sqlite-core/session.js:76:17)
web:dev:     at SQLiteSelectBase._prepare (webpack-internal:///(rsc)/../../node_modules/drizzle-orm/sqlite-core/query-builders/select.js:586:88)
web:dev:     at SQLiteSelectBase.get (webpack-internal:///(rsc)/../../node_modules/drizzle-orm/sqlite-core/query-builders/select.js:604:17)
web:dev:     at getUserByAccount (webpack-internal:///(rsc)/../../node_modules/@auth/drizzle-adapter/lib/sqlite.js:124:18)
web:dev:     at acc.<computed> (webpack-internal:///(rsc)/../../node_modules/next-auth/node_modules/@auth/core/lib/init.js:178:30)
web:dev:     at Module.callback (webpack-internal:///(rsc)/../../node_modules/next-auth/node_modules/@auth/core/lib/actions/callback/index.js:53:39)
web:dev:     at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
web:dev:     at async AuthInternal (webpack-internal:///(rsc)/../../node_modules/next-auth/node_modules/@auth/core/lib/index.js:39:24)
web:dev:     at async Auth (webpack-internal:///(rsc)/../../node_modules/next-auth/node_modules/@auth/core/index.js:126:34)
web:dev: [auth][details]: {}
web:dev:  ⨯ Error [TypeError]: immutable

I've tried to debug adding the debug: true flag to next-auth but there was only not useful information.
From this part web:dev: [auth][details]: {} somehow the data is empty and I think the drizzle adaptor is causing this error.
Actually, I'm new to drizzle too ;-(
How can I solve this error?

"Thank you very much for developing such a fantastic program. However, I have a suggestion. Could we possibly achieve using a local LLM with the OpenAI-like port by
modifying OpenAI's base URL address?"

Can it support SRS/reminder/notification for important data?

for every search, tool may suggest one or more results from database. If there are more than 1 results , is there any way for human/user/tool to prioritise results? (some kind of score).

So that next time, if same search is carried out, results are sorted by its score.

Will there be a Firefox extension anytime soon?

So I had multiple questions in sequence, but i had noticed that my previous prompt answer is coming in next prompt entered,

As you can see when i had asked for timeline i got nothing,
then when i asked for distributed sql i got answer of timeline of evolution,
then when i asked for evolution , i got answer for previous prompt on Distributed SQL
same happens again.

Possible Reasons to Issue

• maybe i had entered new prompt while previous one was processing (but UI was showing no processing like)
• something wrong with Prompt delivery

I tried to deploy the project locally, but when running the bun install command, warning: incorrect peer dependency "[email protected]" appeared. This was due to the two @vitejs/plugin-react": "^4.2.1 and @crxjs /vite-plugin": "^1.0.14 caused by conflicts, how to avoid this?

When I first logged in the only thing I knew was that I can save a webpage and ask question on it but I had no idea of other functions like what does filter do and that purple star thing for bookmark was because of supermemory (lol)

Now that I think its pretty obvious but took me some time for me to figure out
It would be great if we gave the user info about what all the feature at the beginning like a tutorial or maybe a help tab or tool tip explaining what it does

ps: i still dont know what filter does :D

If you type something in the textarea while the response is being stream, after streaming is done, the value is set to ""(empty)

There's a Little inconsistency in border radius under usecases section.
If its intentional you can close this issue and the linked PR to this.

Those inconsistent border radius annoy me a lot 😭

Hello! I am a beginner. When executing the following command, I encountered an issue: "wrangler vectorize create --dimensions=1536 supermem-vector-1 --metric=cosine". During the process, there was an error message:

"✘ [ERROR] A request to the Cloudflare API (/accounts/fec07a08be890e0ad6303b98d06e74ad/vectorize/indexes) failed.

vectorize.not_entitled [code: 1005]"

What does this error message imply? If this issue arises from a mistake on my part, I apologize for any inconvenience caused, and would you kindly assist me in resolving this problem?

the extension and auth text available on sidebar do not provide any functionality

Firefox reader mode simplifies many websites by extracting just the content, the library that does this is open source: https://github.com/mozilla/readability

Page extraction could use this to better target page content. If the backend can accept content directly (Instead of trying to go extract it itself), it's conceivable that the browser extension could use this as a method to extract that content early.

This code is mainly here https://github.com/Dhravya/supermemory/blob/5af20f7b6f2cfceed999dbd9e6244238c82a46b4/apps/cf-ai-backend/src/helper.ts#L82

Currently, we are storing a new vector for each user, in this way:
User -> saves google.com -> first, old stuff is deleted -> vector id is google.com/#supermemory-${userid} (which is then reduced to less than 63 bytes using seededRandom()) -> The page content is chunked and each chunk is saved with it's own ID to the KV. a user: userID metadata is added for retrieval.
User 2 -> saves google.com -> duplicate is made and same after that.

Instead, we want to do it in this way, so that there's no duplicates:
User 1 -> saves google.com -> vector id is google.com (which is reduced to less than 63 bytes using seededrandom()) -> Page content is chunked and each chunk is saved with vectorid-chunkid to the KV. THIS TIME, a metadata with user_${userid}: 1 should be added.

User 2 -> we use the vectorize index to get it by ID
let ids = ["11", "22", "33", "44"]; // all chunk Ids, that we can get by doing list with prefix: seededRandom(url)
const vectors = await env.VECTORIZE_INDEX.getByIds(ids);
If found, simply do env.vectorize_index.upsert with the same vector BUT this time with updated metadata (for each chunk), with the user_${userid2}: 1 added to the json. We don't need to add new documents again. I think we don't even need to update the KV again (since KV is just a lookup of our seededRandom for Url)
NOTE: we also need to change the space adding logic, by adding new keys for each space. like space-userid-spaceid: 1 format.

We would also need to update the retrieval logic in /api/chat which will mostly be the same except this time the filter for spaceid and userid is different.

Provision for labelling data,
Labels should be Hybrid. i.e. text+Number(weightage defined by human/AI/both)

Reference links:

• [hybrid/combinational](teableio/teable#491)
• #36

Just need to ensure that UI is fluid. So users are tempted to label & score data.

Now

Want to build :

• A new ui to enter their email,password along with Continue with Google

Questions -

• Which db are using
• Any inputs from you to build this feature

This is something that I've been wanting to play with for a long time.

However most implementations tend to be based on a browser extension, chrome only, using a proprietary backend or paid service. This project is exciting since the back end is included!!

Are there plans for this to be "one click" self-hostable through docker/K8's, so devs can use it as a solution and build on top of it?

Of course I understand that it could just be forked and modified to fit whatever needs someone might want for it. But improving an existing project is usually better than siloing!

This requires both frontend and backend updates. The goal is to make the chat history feature smooth and user-friendly.

Hey there - excited about this project, so far its much more robust than the other products I've tried.

Is there a discord or slack for folks who want to contribute seriously? Are you open to active development from others, and if so, how do you recommend the community goes about it?

Trying to deploy gives the error:

✘ [ERROR] Running configuration file validation for Pages:

- Configuration file for Pages projects does not support "unsafe"

Its due to the unsafe.binding for ratelimiter

Thanks for the inspiring addon!
Can you add support for Firefox? Thanks.