No description
Find a file
2025-10-16 10:45:00 -07:00
.github ci(workflows): temporarily use php-max version 2025-07-02 13:18:08 -07:00
.tx fix(l10n): reset Transifex translations, update config 2025-03-18 22:50:51 -07:00
appinfo chore: add AI app category 2025-10-06 13:59:02 +02:00
l10n fix(l10n): Update translations from Transifex 2025-10-11 01:05:19 +00:00
lib fix: reset migration versions 2025-04-11 20:14:59 -07:00
LICENSES fix: remove OpenAI logos 2025-03-18 22:51:28 -07:00
src fix: revert "revert outlining key icon" 2025-07-21 17:58:56 -07:00
templates fix: remove text-to-image task processing provider and related settings 2025-03-21 11:53:12 -07:00
tests ci(tests): update bootstrap.php 2025-07-02 13:18:08 -07:00
vendor-bin chore(deps-dev): bump vimeo/psalm from 6.10.0 to 6.10.3 in /vendor-bin/psalm-phar 2025-05-08 16:34:09 -07:00
.eslintrc.cjs docs(reuse): Add SPDX header 2025-01-08 19:47:08 +01:00
.gitattributes docs(reuse): Add SPDX header 2025-01-08 19:47:08 +01:00
.gitignore docs(reuse): Add SPDX header 2025-01-08 19:47:08 +01:00
.l10nignore docs(reuse): Add SPDX header 2025-01-08 19:47:08 +01:00
.nextcloudignore docs(reuse): Add SPDX header 2025-01-08 19:47:08 +01:00
.php-cs-fixer.dist.php docs(reuse): Add SPDX header 2025-01-08 19:47:08 +01:00
AUTHORS.md docs(reuse): Add SPDX header 2025-01-08 19:47:08 +01:00
CHANGELOG.md chore: update CHANGELOG.md 2025-07-22 08:18:17 -07:00
composer.json chore: fix composer.json, update packages 2025-04-11 09:07:49 -07:00
composer.lock chore: fix composer.json, update packages 2025-04-11 09:07:49 -07:00
COPYING it's alive, alive!!! 2022-12-31 17:57:32 +01:00
krankerl.toml docs(reuse): Add SPDX header 2025-01-08 19:47:08 +01:00
makefile fix: rename all instances of OpenAI/LocalAI to watsonx 2025-03-18 22:54:47 -07:00
package-lock.json fix(deps): Fix npm audit 2025-10-12 03:22:03 +00:00
package.json fix: update app text 2025-04-07 13:34:37 -07:00
psalm.xml fix: ignore missing override attribute psalm warnings, fix additional errors 2025-04-10 11:23:51 -07:00
README.md Show REUSE status badge 2025-05-09 11:59:46 +02:00
REUSE.toml fix: rename all instances of OpenAI/LocalAI to watsonx 2025-03-18 22:54:47 -07:00
stylelint.config.cjs docs(reuse): Add SPDX header 2025-01-08 19:47:08 +01:00
vite.config.ts improve vite config 2024-11-18 14:04:49 +01:00

IBM watsonx AI integration in Nextcloud

REUSE status

This app implements the following text generation providers using any freely-available large language model: Free prompt, Summarize, Headline, Context Write, Chat, and Reformulate.

You can connect to the watsonx.ai service provided by IBM or to a self-hosted cluster running IBM Software Hub and related services.

Improve AI task pickup speed

To avoid task processing execution delay, setup at 4 background job workers in the main server (where Nextcloud is installed). The setup process is documented here: https://docs.nextcloud.com/server/latest/admin_manual/ai/overview.html#improve-ai-task-pickup-speed

Ethical AI Rating

Rating for Text generation via IBM watsonx.ai: 🟠

Positive:

  • The provided foundation models are freely available, and thus can be ran on-premises

Negative:

  • The software for training and inference of models is proprietary, limiting modifications to the API or other functionality
  • Some foundation models are trained on data that is not freely available, limiting the ability to fine tune them

Learn more about the Nextcloud Ethical AI Rating in our blog.

Limitations

Warning

This app is still in early development and has only been tested with IBM watsonx.ai as a Service. The following list details some missing features that may be added in a future release.

  • Support for agency features (i.e. IBM watsonx as chat with tools provider)
  • Support for additional models without complete API functionality (see: https://www.ibm.com/watsonx/developer/get-started/models/)
  • Support for IBM Cloud Pak for Data Platform API (for identity management on self-hosted instances)
  • Support for more than 100 models deployed in IBM watsonx.ai
  • Ability to select an IBM Cloud location from a dropdown list (as a workaround, enter the location's API URL manually)

🔧 Configuration

Admin settings

There is a "Artificial intelligence" admin settings section where you can:

  • Choose whether you use an IBM-hosted watsonx.ai instance or another remote service
  • Set a global API key and cloud resource for the Nextcloud instance
  • Configure default models and quota settings

Personal settings

There is a "Artificial intelligence" personal settings section to let users set their personal API key and cloud resources. Users can also see their quota information there.