[PR #15569] [MERGED] launch: add hermes #77506

Closed
opened 2026-05-05 10:10:38 -05:00 by GiteaMirror · 0 comments
Owner

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/15569
Author: @ParthSareen
Created: 4/14/2026
Status: Merged
Merged: 4/15/2026
Merged by: @ParthSareen

Base: mainHead: parth-launch-hermes


📝 Commits (10+)

  • ecec882 launch: add hermes to launch
  • 083f2dc launch: fix launcher state lint
  • 793fff9 update go.mod
  • 25e2182 fix doc
  • 5c1b625 launch: stabilize hermes integration flow
  • 0eb2d58 Prompt before installing Hermes
  • 4c70714 launch: trim hermes wsl tests
  • 827f2a2 Skip Hermes setup wizard and prompt for messaging
  • 5c8d35c Fix launcher integration menu order
  • c91d84f Restart Hermes gateway after config updates

📊 Changes

12 files changed (+3036 additions, -110 deletions)

View changed files

📝 cmd/launch/command_test.go (+3 -0)
cmd/launch/hermes.go (+962 -0)
cmd/launch/hermes_test.go (+1236 -0)
📝 cmd/launch/integrations_test.go (+30 -21)
📝 cmd/launch/launch.go (+199 -42)
📝 cmd/launch/launch_test.go (+491 -0)
📝 cmd/launch/models.go (+17 -3)
📝 cmd/launch/registry.go (+17 -3)
📝 cmd/tui/tui.go (+37 -37)
📝 cmd/tui/tui_test.go (+39 -3)
📝 docs/integrations/hermes.mdx (+4 -0)
📝 go.mod (+1 -1)

📄 Description

First pass at hermes agent integration.

Includes ollama launch hermes with install and onboarding baked in.

What changed:

  • Added Hermes to the launch registry, help text, tests, and launcher UI.
  • Added a narrow ManagedSingleModel flow in launcher code so Hermes can:
    • pick one Ollama-backed primary model through launch
    • persist launch-owned config
    • report current model / onboarding state back to launcher
  • Implemented Hermes-specific config management:
    • writes Hermes’ primary model + Ollama /v1 endpoint into ~/.hermes/config.yaml
    • uses a launch-owned provider key (ollama-launch)
    • preserves unrelated Hermes config
    • migrates older launch-managed Ollama aliases
    • keeps model discovery tied to the launch-resolved Ollama host
  • Added Hermes install/setup behavior:
    • prompt-driven install on macOS/Linux
    • WSL-first behavior on Windows
    • optional messaging gateway setup prompt on interactive launch
    • gateway restart after config changes if Hermes’ messaging runtime is already running
  • Updated launcher state/TUI behavior so Hermes shows up in the interactive launcher and respects current model/install status.
  • Added a small Hermes docs entry for ollama launch hermes.

Things To Look For

  • cmd/launch/launch.go: the new ManagedSingleModel path is generic enough for Hermes without making the rest of launcher logic harder to follow.
  • cmd/launch/hermes.go: the config ownership rules are sane, especially around providers.ollama-launch, legacy alias migration, and stale-config detection.
  • Windows/WSL behavior: install handoff, runtime path selection, and whether config writes line up with the Hermes binary that actually runs.
  • Gateway/messaging behavior: optional preflight prompt on interactive launch and restart-after-config-change behavior if the gateway is already running.
  • TUI/menu behavior: Hermes is visible in launcher ordering, but still sits under More... because we only pin the first 3 integrations.

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/ollama/ollama/pull/15569 **Author:** [@ParthSareen](https://github.com/ParthSareen) **Created:** 4/14/2026 **Status:** ✅ Merged **Merged:** 4/15/2026 **Merged by:** [@ParthSareen](https://github.com/ParthSareen) **Base:** `main` ← **Head:** `parth-launch-hermes` --- ### 📝 Commits (10+) - [`ecec882`](https://github.com/ollama/ollama/commit/ecec882802f52651da15f3601a9d9e80c247832f) launch: add hermes to launch - [`083f2dc`](https://github.com/ollama/ollama/commit/083f2dc52244d3edc620f6e90e575068eb59016f) launch: fix launcher state lint - [`793fff9`](https://github.com/ollama/ollama/commit/793fff9cc062d74a9f707faf573b6e87e34d2827) update go.mod - [`25e2182`](https://github.com/ollama/ollama/commit/25e218277df27aa1347b358073d5f8176d5c3d53) fix doc - [`5c1b625`](https://github.com/ollama/ollama/commit/5c1b6255f2e2d71c8320870778f1e2c16d3381b8) launch: stabilize hermes integration flow - [`0eb2d58`](https://github.com/ollama/ollama/commit/0eb2d58c14eaab19f2ab36a8b334af34887c9155) Prompt before installing Hermes - [`4c70714`](https://github.com/ollama/ollama/commit/4c7071422281ad7533418aa97fbac315838a0a04) launch: trim hermes wsl tests - [`827f2a2`](https://github.com/ollama/ollama/commit/827f2a216a0f34907d5f0adc4919a2bcb5fdd190) Skip Hermes setup wizard and prompt for messaging - [`5c8d35c`](https://github.com/ollama/ollama/commit/5c8d35c66edb53b62010b72ed6abfebbce790ce8) Fix launcher integration menu order - [`c91d84f`](https://github.com/ollama/ollama/commit/c91d84fe77c3739ac75498e6a76ef0718bfae378) Restart Hermes gateway after config updates ### 📊 Changes **12 files changed** (+3036 additions, -110 deletions) <details> <summary>View changed files</summary> 📝 `cmd/launch/command_test.go` (+3 -0) ➕ `cmd/launch/hermes.go` (+962 -0) ➕ `cmd/launch/hermes_test.go` (+1236 -0) 📝 `cmd/launch/integrations_test.go` (+30 -21) 📝 `cmd/launch/launch.go` (+199 -42) 📝 `cmd/launch/launch_test.go` (+491 -0) 📝 `cmd/launch/models.go` (+17 -3) 📝 `cmd/launch/registry.go` (+17 -3) 📝 `cmd/tui/tui.go` (+37 -37) 📝 `cmd/tui/tui_test.go` (+39 -3) 📝 `docs/integrations/hermes.mdx` (+4 -0) 📝 `go.mod` (+1 -1) </details> ### 📄 Description First pass at hermes agent integration. Includes `ollama launch hermes` with install and onboarding baked in. What changed: - Added Hermes to the launch registry, help text, tests, and launcher UI. - Added a narrow `ManagedSingleModel` flow in launcher code so Hermes can: - pick one Ollama-backed primary model through launch - persist launch-owned config - report current model / onboarding state back to launcher - Implemented Hermes-specific config management: - writes Hermes’ primary model + Ollama `/v1` endpoint into `~/.hermes/config.yaml` - uses a launch-owned provider key (`ollama-launch`) - preserves unrelated Hermes config - migrates older launch-managed Ollama aliases - keeps model discovery tied to the launch-resolved Ollama host - Added Hermes install/setup behavior: - prompt-driven install on macOS/Linux - WSL-first behavior on Windows - optional messaging gateway setup prompt on interactive launch - gateway restart after config changes if Hermes’ messaging runtime is already running - Updated launcher state/TUI behavior so Hermes shows up in the interactive launcher and respects current model/install status. - Added a small Hermes docs entry for `ollama launch hermes`. **Things To Look For** - `cmd/launch/launch.go`: the new `ManagedSingleModel` path is generic enough for Hermes without making the rest of launcher logic harder to follow. - `cmd/launch/hermes.go`: the config ownership rules are sane, especially around `providers.ollama-launch`, legacy alias migration, and stale-config detection. - Windows/WSL behavior: install handoff, runtime path selection, and whether config writes line up with the Hermes binary that actually runs. - Gateway/messaging behavior: optional preflight prompt on interactive launch and restart-after-config-change behavior if the gateway is already running. - TUI/menu behavior: Hermes is visible in launcher ordering, but still sits under `More...` because we only pin the first 3 integrations. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-05-05 10:10:38 -05:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#77506