[GH-ISSUE #66] [Enhancement]: intergrations for the best open source model as of today #34

New Issue

2026-06-06T22:08:09-04:00

yindo commented

2026-06-06 22:08:09 -04:00

Originally created by @s-b-repo on GitHub (Oct 25, 2025).
Original GitHub issue: https://github.com/vxcontrol/pentagi/issues/66

Originally assigned to: @asdek on GitHub.

Target Component

External Integrations (LLM/Search APIs)

Enhancement Description

https://www.kimi.com

Technical Details

N/A

Designs and Mockups

N/A

Alternative Solutions

none

Verification

I have checked that this enhancement hasn't been already proposed
This enhancement aligns with PentAGI's goal of autonomous penetration testing
I have considered the security implications of this enhancement
I have provided clear use cases and benefits

Originally created by @s-b-repo on GitHub (Oct 25, 2025). Original GitHub issue: https://github.com/vxcontrol/pentagi/issues/66 Originally assigned to: @asdek on GitHub. ### Target Component External Integrations (LLM/Search APIs) ### Enhancement Description https://www.kimi.com ### Technical Details N/A ### Designs and Mockups N/A ### Alternative Solutions none ### Verification - [x] I have checked that this enhancement hasn't been already proposed - [x] This enhancement aligns with PentAGI's goal of autonomous penetration testing - [x] I have considered the security implications of this enhancement - [x] I have provided clear use cases and benefits

yindo added the enhancement label 2026-06-06 22:08:09 -04:00

yindo closed this issue

2026-06-06 22:08:09 -04:00

yindo commented

2026-06-06 22:08:10 -04:00

@asdek commented on GitHub (Oct 25, 2025):

Hello @s-b-repo, thank you for your request,

PentAGI supports Kimi K2 LLM via OpenRouter and DeepInfra, and I also added a configuration for Moonshot API in f2f51dd.

You can try this using the following environment variables:

LLM_SERVER_URL=https://api.moonshot.ai/v1
LLM_SERVER_KEY=your_api_key
LLM_SERVER_MODEL=kimi-k2-0905-preview
LLM_SERVER_CONFIG_PATH=/opt/pentagi/conf/moonshot.provider.yml

Before running, execute docker compose pull and then docker compose up -d

Regarding the web-search feature, I studied the documentation on how this can be used through the API, and it seems that such behavior is too unique for this LLM provider. OpenAI implements this through Response API, Anthropic uses a different type of function, Google Gemini requires a full JSON schema for built-in functions, etc.

Overall, I don't see a great need to use this capability, as PentAGI already contains a sufficient number of search engines, namely: Built-in browser via scraper for gathering the latest information from web sources, advanced search APIs including Tavily, Traversaal, Perplexity, DuckDuckGo, Google Custom Search, and Searxng for comprehensive information gathering. The built-in browser and DuckDuckGo are enabled by default, Searxng needs to be installed separately, and the rest require API keys, but their usage costs don't differ significantly from Moonshot, so this change is not currently planned.

@asdek commented on GitHub (Oct 25, 2025): Hello @s-b-repo, thank you for your request, PentAGI supports Kimi K2 LLM via OpenRouter and DeepInfra, and I also added a [configuration](https://github.com/vxcontrol/pentagi/blob/master/examples/configs/moonshot.provider.yml) for Moonshot API in [f2f51dd](https://github.com/vxcontrol/pentagi/commit/f2f51dd5e262fc1bd671a1df24887265f789f501). You can try this using the following environment variables: ``` LLM_SERVER_URL=https://api.moonshot.ai/v1 LLM_SERVER_KEY=your_api_key LLM_SERVER_MODEL=kimi-k2-0905-preview LLM_SERVER_CONFIG_PATH=/opt/pentagi/conf/moonshot.provider.yml ``` Before running, execute `docker compose pull` and then `docker compose up -d` Regarding the web-search feature, I studied the [documentation](https://platform.moonshot.ai/docs/guide/use-web-search.en-US) on how this can be used through the API, and it seems that such behavior is too unique for this LLM provider. OpenAI implements this through Response API, Anthropic uses a different type of function, Google Gemini requires a full JSON schema for built-in functions, etc. Overall, I don't see a great need to use this capability, as PentAGI already contains a sufficient number of search engines, namely: Built-in browser via [scraper](https://hub.docker.com/r/vxcontrol/scraper) for gathering the latest information from web sources, advanced search APIs including [Tavily](https://tavily.com), [Traversaal](https://traversaal.ai), [Perplexity](https://www.perplexity.ai), [DuckDuckGo](https://duckduckgo.com/), [Google Custom Search](https://programmablesearchengine.google.com/), and [Searxng](https://searxng.org) for comprehensive information gathering. The built-in browser and DuckDuckGo are enabled by default, Searxng needs to be installed separately, and the rest require API keys, but their usage costs don't differ significantly from Moonshot, so this change is not currently planned.

yindo referenced this issue

2026-06-06 22:09:19 -04:00

[PR #34] [MERGED] chore(deps): bump golang.org/x/net from 0.36.0 to 0.38.0 in /backend in the go_modules group across 1 directory #132

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: vxcontrol/pentagi#34