There should be a toggle button to switch back and forth between emails shown and hidden in the tables and charts

Currently the userId for a specific request is set through the X-User-Id header, like this:

const configuration = new Configuration({
  apiKey: process.env.OPENAI_API_KEY, 
  basePath: "https://api.openai.withlogging.com/v1",  
  baseOptions: { 
    headers: {
      "X-Api-Key": `Bearer ${process.env.LLM_REPORT_API_KEY}`, 
      "X-User-Id": `[email protected]`,  // user Id
    },
  }
});

This means that, if different requests need to be associated to different users, the request header needs to be dynamic and change on each request (and therefore we need to create a new Configuration object on each request).

However, OpenAI API already accepts a user field in the body of the request, which has a similar purpose.

const completion = await openai.createChatCompletion({
  model: "gpt-3.5-turbo",
  user: "userid",
});

It would be handy to optionally get the userId for LLM Report from this field of the request body, rather than the X-User-Id header. This would avoid having to set new headers for each request.

Looks like there is some weird overflow scrolling issue on the /logs page

Great app! Build an official image and push it to a docker registry like Dockerhub or Github package would be awesome, so the ones that selfhost would be always up to date :)

on the /openai tab, going to detailed analytics causes computer to lag

most likely due to minute chart

possible solution:

• show minute chart but have a date dropdown to select day. This will have less load on the chart
• another solution: so default group data by hour, and if you select a specific day, group by minute

Noticed this issue occurring once a day now.

Here's the 2 error messages associated with the issue:

Error: fail to memory allocation
at /app/node_modules/vscode-oniguruma/release/main.js:1:3101
at new u (/app/node_modules/vscode-oniguruma/release/main.js:1:3150)
at Object.n.createOnigScanner (/app/node_modules/vscode-oniguruma/release/main.js:1:4984)
at Object.createOnigScanner (/app/node_modules/shiki/dist/index.js:1932:34)
at g.createOnigScanner (/app/node_modules/vscode-textmate/release/main.js:1:4532)
at new m (/app/node_modules/vscode-textmate/release/main.js:1:47463)
at g.compile (/app/node_modules/vscode-textmate/release/main.js:1:46704)
at g.compileAG (/app/node_modules/vscode-textmate/release/main.js:1:47238)
at u.compileAG (/app/node_modules/vscode-textmate/release/main.js:1:38683)
at l (/app/node_modules/vscode-textmate/release/main.js:1:22586)

Error: \(([^\(\)]|(\(([^\(\)]|\([^\(\)]*\))*\)))*\))|(\[([^\[\]]|(\[([^\[\]]|\[[^\[\]]*\])*\]))*\]))([^=<>]|=[^<]|\<\s*(((const\s+)?[_$[:alpha:]])|(\{([^\{\}]|(\{([^\{\}]|\{[^\{\}]*\})*\}))*\})|(\(([^\(\)]|(\(([^\(\)]|\([^\(\)]*\))*\)))*\))|(\[([^\[\]]|(\[([^\[\]]|\[[^\[\]]*\])*\]))*\]))([^=<>]|=[^<]|\<\s*(((const\s+)?[_$[:alpha:]])|(\{([^\{\}]|(\{([^\{\}]|\{[^\{\}]*\})*\}))*\})|(\(([^\(\)]|(\(([^\(\)]|\([^\(\)]*\))*\)))*\))|(\[([^\[\]]|(\[([^\[\]]|\[[^\[\]]*\])*\]))*\]))([^=<>]|=[^<])*\>)*\>)*>\s*)?                                                                                 # typeparameters
\(\s*(\/\*([^\*]|(\*[^\/]))*\*\/\s*)*(([_$[:alpha:]]|(\{([^\{\}]|(\{([^\{\}]|\{[^\{\}]*\})*\}))*\})|(\[([^\[\]]|(\[([^\[\]]|\[[^\[\]]*\])*\]))*\])|(\.\.\.\s*[_$[:alpha:]]))([^()\'\"\`]|(\(([^\(\)]|(\(([^\(\)]|\([^\(\)]*\))*\)))*\))|(\'([^\'\\]|\\.)*\')|(\"([^\"\\]|\\.)*\")|(\`([^\`\\]|\\.)*\`))*)?\)   # parameters
(\s*:\s*([^<>\(\)\{\}]|\<([^<>]|\<([^<>]|\<[^<>]+\>)+\>)+\>|\([^\(\)]+\)|\{[^\{\}]+\})+)?                                                                        # return type
\s*=>                                                                                               # arrow operator
)
))
))x
at /app/node_modules/vscode-oniguruma/release/main.js:1:3101
at new u (/app/node_modules/vscode-oniguruma/release/main.js:1:3150)
at Object.n.createOnigScanner (/app/node_modules/vscode-oniguruma/release/main.js:1:4984)
at Object.createOnigScanner (/app/node_modules/shiki/dist/index.js:1932:34)
at g.createOnigScanner (/app/node_modules/vscode-textmate/release/main.js:1:4532)
at new m (/app/node_modules/vscode-textmate/release/main.js:1:47463)
at g._resolveAnchors (/app/node_modules/vscode-textmate/release/main.js:1:47329)
at g.compileAG (/app/node_modules/vscode-textmate/release/main.js:1:46972)
at u.compileAG (/app/node_modules/vscode-textmate/release/main.js:1:38683)
at l (/app/node_modules/vscode-textmate/release/main.js:1:22586)
[Error: An error occurred in the Server Components render. The specific message is omitted in production builds to avoid leaking sensitive details. A digest property is included on this error instance which may provide additional details about the nature of the error.] {
digest: '3060834878'
}

Based off the error messages, it's related to the shiki dependency which is used for syntax highlighting of this code snippet. If someone knows why this is happening that would be really helpful!

   "shiki": "^0.14.2"

Origin file https://github.com/dillionverma/llm.report/blob/main/lib/markdown-code.ts
Used in this file: https://github.com/dillionverma/llm.report/blob/main/app/(dashboard)/logs/logs-onboarding.tsx

Possible solutions:

• Add a try catch so it doesn't break the entire site?
• Remove shiki as dependency overall
• Try upgrading to latest version hoping its fixed 🤞

When I enter the sess-key, I do not get the expected page.

Caching is awesome to save on tokens, but there are few scenarios where someone might want to opt out of it.

E.g. allowing to re-generate a response if the user doesn't like it. In this case the prompt is the same as a previous request (so the response would be taken from the cache) but we actually want to obtain a fresh new result.

I would then suggest to allow to opt out of caching on a request basis. Maybe using a specific header for this?

https://blog.railway.app/p/automated-postgresql-backups

need to pass user id as url query param to openai api requests like so:

https://api.openai.com/v1/usage?date=2023-08-23&user_public_id=user-ji6LpnyrRSo6

Inspect the network requests on https://platform.openai.com/account/usage website to see how the requests / responses are sent

Need to update the following file to get started and then work backward throughout the app - https://github.com/dillionverma/llm.report/blob/main/lib/services/openai.ts

If a user has a specific log - i.e.

https://llm.report/logs/chatcmpl-7vWp52MLmQpMaelMnLH0H30

the app should open up the log drawer and the details for that specific log

on the /logs page there should be a way to filter by model

Add langchain code snippets for python / js

Here's example of js snippet which works

Would need to add a new snippet to the in-app installation instructions

Hey,

I did try running your website locally by filling all the fields of the .env.local file.
It seems however that something is wrong, the posthog api key is not considered valid (i am not authorized) and I don't get if it's because I'm missing some flags in my posthog project or what

also it's not clear to me what the withlogging.com website has to do with you, how the api behind it works etc.
I can't find it in the code so I don't understand what happens exactly when I fetch a completion by using your domain rather than the original openai one.

would you mind guiding me on how to use all the features of your saas in a local environment for testing purposes?

Pricing page should redirect to /login if user not logged in

Relevant function to change

Drawer card should change if user is a paying user