FastGPT API Failed to process request

sauterp

Steps to reproduce:
Use this script

#!/usr/bin/env sh

query=$@

response=$(curl --silent \
    -H "Authorization: Bot $KAGI_API_TOKEN" \
    -H "Content-Type: application/json" \
    --data "{\"query\": \"$query\"}" \
    https://kagi.com/api/v0/fastgpt)

echo $response | jq
answer=$(echo $response | jq '.data.output')
echo -e $answer

now execute the script with a query like fastgpt who invented unix

Expected behavior:
This same script used to work a few minutes ago, but now it fails. I did quite a lot of requests before. Am I being rate limited?

Unfortunately I get this:

{
  "meta": {
    "id": "0ab037b3-fe17-43ed-9a54-f0f390015abc",
    "node": "europe-west4",
    "ms": 106
  },
  "data": {
    "output": null,
    "tokens": 0
  },
  "error": [
    {
      "code": 200,
      "msg": "Failed to process request",
      "ref": null
    }
  ]
}

Debug info:
verbose curl output:

./executable_fastgpt who invented unix
*   Trying 34.111.242.115:443...
* Connected to kagi.com (34.111.242.115) port 443 (#0)
* ALPN: offers h2
* ALPN: offers http/1.1
*  CAfile: /etc/pki/tls/certs/ca-bundle.crt
*  CApath: none
* TLSv1.0 (OUT), TLS header, Certificate Status (22):
} [5 bytes data]
* TLSv1.3 (OUT), TLS handshake, Client hello (1):
} [512 bytes data]
* TLSv1.2 (IN), TLS header, Certificate Status (22):
{ [5 bytes data]
* TLSv1.3 (IN), TLS handshake, Server hello (2):
{ [122 bytes data]
* TLSv1.2 (IN), TLS header, Finished (20):
{ [5 bytes data]
* TLSv1.2 (IN), TLS header, Supplemental data (23):
{ [5 bytes data]
* TLSv1.3 (IN), TLS handshake, Encrypted Extensions (8):
{ [15 bytes data]
* TLSv1.3 (IN), TLS handshake, Certificate (11):
{ [4029 bytes data]
* TLSv1.3 (IN), TLS handshake, CERT verify (15):
{ [264 bytes data]
* TLSv1.3 (IN), TLS handshake, Finished (20):
{ [52 bytes data]
* TLSv1.2 (OUT), TLS header, Finished (20):
} [5 bytes data]
* TLSv1.3 (OUT), TLS change cipher, Change cipher spec (1):
} [1 bytes data]
* TLSv1.2 (OUT), TLS header, Supplemental data (23):
} [5 bytes data]
* TLSv1.3 (OUT), TLS handshake, Finished (20):
} [52 bytes data]
* SSL connection using TLSv1.3 / TLS_AES_256_GCM_SHA384
* ALPN: server accepted h2
* Server certificate:
*  subject: CN=*.kagi.com
*  start date: Apr 17 13:34:22 2023 GMT
*  expire date: Jul 16 13:34:21 2023 GMT
*  subjectAltName: host "kagi.com" matched cert's "kagi.com"
*  issuer: C=US; O=Let's Encrypt; CN=R3
*  SSL certificate verify ok.
* Using HTTP2, server supports multiplexing
* Copying HTTP/2 data in stream buffer to connection buffer after upgrade: len=0
* TLSv1.2 (OUT), TLS header, Supplemental data (23):
} [5 bytes data]
* TLSv1.2 (OUT), TLS header, Supplemental data (23):
} [5 bytes data]
* TLSv1.2 (OUT), TLS header, Supplemental data (23):
} [5 bytes data]
* h2h3 [:method: POST]
* h2h3 [:path: /api/v0/fastgpt]
* h2h3 [:scheme: https]
* h2h3 [:authority: kagi.com]
* h2h3 [user-agent: curl/7.85.0]
* h2h3 [accept: */*]
* h2h3 [content-type: application/json]
* h2h3 [content-length: 30]
* TLSv1.2 (OUT), TLS header, Supplemental data (23):
} [5 bytes data]
> POST /api/v0/fastgpt HTTP/2
> Host: kagi.com
> user-agent: curl/7.85.0
> accept: */*
> content-type: application/json
> content-length: 30
>
* TLSv1.2 (OUT), TLS header, Supplemental data (23):
} [5 bytes data]
* We are completely uploaded and fine
* TLSv1.2 (IN), TLS header, Supplemental data (23):
{ [5 bytes data]
* TLSv1.3 (IN), TLS handshake, Newsession Ticket (4):
{ [267 bytes data]
* TLSv1.3 (IN), TLS handshake, Newsession Ticket (4):
{ [267 bytes data]
* old SSL session ID is stale, removing
* TLSv1.2 (IN), TLS header, Supplemental data (23):
{ [5 bytes data]
* TLSv1.2 (OUT), TLS header, Supplemental data (23):
} [5 bytes data]
* TLSv1.2 (IN), TLS header, Supplemental data (23):
{ [5 bytes data]
* TLSv1.2 (IN), TLS header, Supplemental data (23):
{ [5 bytes data]
< HTTP/2 200
< x-request-id: 9f87e5f7-c824-4a7d-bbdf-e19b7faed389
< cache-control: no-store
< content-type: application/json
< content-length: 187
< date: Tue, 06 Jun 2023 07:34:20 GMT
< via: 1.1 google
<
* TLSv1.2 (IN), TLS header, Supplemental data (23):
{ [5 bytes data]
* TLSv1.2 (IN), TLS header, Supplemental data (23):
{ [5 bytes data]
* Connection #0 to host kagi.com left intact
{
  "meta": {
    "id": "9f87e5f7-c824-4a7d-bbdf-e19b7faed389",
    "node": "europe-west4",
    "ms": 105
  },
  "data": {
    "output": null,
    "tokens": 0
  },
  "error": [
    {
      "code": 200,
      "msg": "Failed to process request",
      "ref": null
    }
  ]
}

anijatsu

It does seem like something broke since the last update (1591).
Instant Answers, Universal Summarizer don't work, but the FastGPT accessed through the page (https://labs.kagi.com/fastgpt) does.

anijatsu

Steps to reproduce:
Make any Kagi query and try to use Instant Answer, or
try to summarize any website through Kagi or Universal Summ.'s own website (https://kagi.com/summarizer).
Yet, observe that fastGPT's own page continues to work: https://labs.kagi.com/fastgpt

Expected behavior:
The functions continue to work.

Debug info:
Connected to: EUROPE-WEST2

Image/Video:

pieterclaerhout

Steps to reproduce:

We're trying to get summaries using a POST of a JSON body as instructed in the documentation:

https://help.kagi.com/kagi/api/summarizer.html (under "POST request with JSON body, song lyrics, Cecil engine").

However, when we execute the example with our token, we get:

curl -X "POST" "https://kagi.com/api/v0/summarize" \
     -H 'Authorization: Bot <token>' \
     -H 'Content-Type: application/json; charset=utf-8' \
     -d $'{
  "url": "https://genius.com/Brothertiger-mainsail-lyrics",
  "engine": "cecil"
}'

The response returns:

{
  "meta": {
    "id": "26a73ab4-e15e-413f-a585-6bfd0c7f3619",
    "node": "europe-west4",
    "ms": 6
  },
  "data": {
    "output": null,
    "tokens": 0
  },
  "error": [
    {
      "code": 200,
      "msg": "Sorry, a problem occurred while processing your request. Please try again later.",
      "ref": null
    }
  ]
}

It's a bit weird that we get code "200" along with an error message (as 200 usually means that everything went fine).

Also, the error message isn't really helpful and doesn't give us any clue on what the problem might be.

Expected behavior:

We would expect a summary to be returned without any errors (especially since this is an example from the documentation).

In the case of an error, a message indicating what the problem and solution are would be very helpful.

Debug info:

cURL running on macOS and / or Linux

Node: europe-west4

z64

Hi there,

Currently out AI services in the European regions are experiencing some issues that we're looking into.

The "200" is from our own error code index, and not an HTTP status code. I see that we need to publish our list of these in our docs; will see to that as well.

z64

Hi all, we've deployed some tweaks to try and improve API & AI stability. Please feel free to update if things are working for you now, or if you're still experiencing this issue. Thanks!

pieterclaerhout

Hi there, the summarize endpoint still gives me the same error message 🙁

sauterp

@z64 thank you, the fastgpt endpoint works again for me.

sauterp

Now the API times out 🙁 I can send queries without error, but I don't get an answer.

sauterp

sauterp works again

pieterclaerhout

Any update on the summarize API? It still doesn't work for me…

z64

Hi pieterclaerhout - we've been making progressive improvements to API stability over the past 48hrs or so, and have seen a majority reduction in reports - if you are still experiencing issues, could you confirm that your last post is still a example of a failing request, and if possible, provide a few new request IDs that we can trace through our backend? They are the meta.id values.

pieterclaerhout

@z64 I've tested again and now it seems to be working again 🙂