1 install Hermes,
In tutorials he states "Open herems app, go to settings,"
follow everything from tech with tim, to install Hermes I'll put install steps here, choose options from video After install you can run Hermes Desktop
This is the curl command,
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
2 run Hermes desktop
https://hermes-agent.nousresearch.com/docs/user-guide/desktop Figured out for hermes desktop just run
hermes desktop
Now we can continue following the tutorial
This wasn't the right hermes desktop
3 chatgpt answer
Chat says, run
curl http://127.0.0.1:11434/v1/models
and
hermes model
which i did above, and it says Choose Custome Endpoint
4 chatgpt answer-2
chat says chooseCustom endpoinit
then for API base URL [e.g. https://api.example.com/v1]: enter
http://127.0.0.1:11434/v1
This is the endpoint for Ollama, but i dont know where this info about endpoint for ollama is
Also, after pasting endpoint, enter NOTHING for API key
4 chatgpt answer-3
STUCK AGAIN, why did it tell me choose 2 for completions, This could be a problem for tool ahead
I chose option 2 chat completions
then choose option 1. gemma4e-64k:latest
then entered context length
64000
because we changed it with the Variant Modelfile previously
5 chatgpt answer-3
STUCK AGAIN, why did it tell me choose 2 for completions, This could be a problem for tool ahead
So i asked chat this 2 part question,
I typed 2, chose 2, but I'm gonna want to use this for tools and agent, so i'm guessing agents and tools won't work with this option,
so i'll probably have change this, but regarless, let's just get anything working now, i just choose option 2 chat completions,
then i selected 1. for gemma4e-64k, then i entered 64000 (because i made a Modelfile that changed that variant to
64000)
and it responded with this
Select API compatibility mode:
1. Auto-detect [current]
Use Hermes URL heuristics; best for standard OpenAI-compatible endpoints.
2. Chat Completions
Use /chat/completions for standard OpenAI-compatible servers.
3. Responses / Codex
Use /responses for Codex-compatible tool-calling backends.
4. Anthropic Messages
Use /v1/messages for Anthropic-compatible endpoints.
Choice [1-4, Enter to keep current/detected]: 2
API mode: chat_completions
Available models:
1. gemma4e-64k:latest
2. gemma4:e4b
Select model [1-2] or type name: 1
Context length in tokens [leave blank for auto-detect]: 64000
Display name [Local (127.0.0.1:11434)]:
so now what
chat response to above question,
for Display name [Local (127.0.0.1:11434)]:
type
Ollama Gemma 4 64K
OR, " Or simply press Enter to keep the default name. Either works."
I'M JUST PRESSING ENTER
chose mediam reasoning, it closes hemres in terminal
6 chatgpt answer-5
I fucking hope this the last step, i choose mediam now it exited and i don't know what to do
I gave this to chat
i choose mediam now it exited and i don't know what to do now, how to i now connect hermes desktop to local enpoint?
It told me to run
hermes model
And asked me for screen shot of hermes panel, i gave it this screent shot
screen shot shows the options in termainl for the command hermes model hermes model, and hermes setting
YESSSSSS We got it, i it's working, i had to open ollama then click heremes desktop which copied the command ollama launch hermes-desktop to clipboard i pasted it into terminal, now the chat and tools are working in heremes desktop like in this video
7 FIX
Click hermes option in ollama to the command,
ollama launch hermes-desktop
paste in terminal, start hermes