Skip to content

feat(agent): add lmstudio integration#17061

Closed
rugvedS07 wants to merge 5 commits into
NousResearch:mainfrom
rugvedS07:feat/lmstudio-integration
Closed

feat(agent): add lmstudio integration#17061
rugvedS07 wants to merge 5 commits into
NousResearch:mainfrom
rugvedS07:feat/lmstudio-integration

Conversation

@rugvedS07

@rugvedS07 rugvedS07 commented Apr 28, 2026

Copy link
Copy Markdown
Contributor

What does this PR do?

This PR adds LM Studio as a first class provider, allowing richer experience with JIT loading with higher context length, reasoning capabilities and supporting auth mode.

Type of Change

  • 🐛 Bug fix (non-breaking change that fixes an issue)
  • ✨ New feature (non-breaking change that adds functionality)
  • 🔒 Security fix
  • 📝 Documentation update
  • ✅ Tests (adding or improving test coverage)
  • ♻️ Refactor (no behavior change)
  • 🎯 New skill (bundled or hub)

Changes Made

  • Adds LM Studio as a first-class provider with aliases, env vars, and a no-auth fallback so local users aren’t blocked.
  • Auto-detects locally loaded models via /api/v1/models, and auto-loads model with at least 64K context using /api/v1/models/load endpoint
  • Maps reasoning settings to LM Studio’s supported “effort” levels and skips unsupported values to prevent errors.
  • Tracks per-model context length and ensures models are loaded at runtime; updates the live context display immediately after load
  • Adds tests for all of the above

How to Test

  1. Onboard hermes with LM Studio as a model provider
  2. Try using /model slash command and /reasoning slash command
  3. You can see reasoning_effort being populated in LM Studio requests

Checklist

Code

  • I've read the Contributing Guide
  • My commit messages follow Conventional Commits (fix(scope):, feat(scope):, etc.)
  • I searched for existing PRs to make sure this isn't a duplicate
  • My PR contains only changes related to this fix/feature (no unrelated commits)
  • I've run pytest tests/ -q and all tests pass
  • I've added tests for my changes (required for bug fixes, strongly encouraged for features)
  • I've tested on my platform: macOS 15.7

Documentation & Housekeeping

  • I've updated relevant documentation (README, docs/, docstrings)
  • I've updated cli-config.yaml.example if I added/changed config
  • I've considered cross-platform impact (Windows, macOS) per the compatibility guide

Screenshots / Logs

Screenshot 2026-04-27 at 12 05 24 PM
image
image
image
image
image
Screenshot 2026-04-27 at 12 19 00 PM

@rugvedS07 rugvedS07 marked this pull request as ready for review April 28, 2026 15:09
@alt-glitch alt-glitch added type/feature New feature or request P3 Low — cosmetic, nice to have comp/agent Core agent loop, run_agent.py, prompt builder comp/cli CLI entry point, hermes_cli/, setup wizard comp/gateway Gateway runner, session dispatch, delivery comp/tui Terminal UI (ui-tui/ + tui_gateway/) labels Apr 28, 2026
@kshitijk4poor kshitijk4poor mentioned this pull request Apr 28, 2026
Merged
6 tasks
kshitijk4poor added a commit that referenced this pull request Apr 28, 2026
@kshitijk4poor
fix: follow-up for salvaged PR #17061
bced623 
- Fix TUI test: relax target_model assertion to avoid module-cache flake
  (pre-existing failure on main, test now asserts the stable part of the call)
- Remove unnecessary config version bump 22→23: runtime already handles
  no-auth via LMSTUDIO_NOAUTH_PLACEHOLDER in resolve_api_key_provider_credentials
- Revert runtime_provider.py base_url precedence change: the original
  contract (saved config.base_url > env var > default) must be preserved
  for all api_key providers, not just lmstudio. LM Studio works correctly
  with the original logic since resolve_api_key_provider_credentials
  already incorporates LM_BASE_URL into creds["base_url"].
- AUTHOR_MAP: added rugved@lmstudio.ai → rugvedS07
kshitijk4poor added a commit that referenced this pull request Apr 28, 2026
@kshitijk4poor
fix: follow-up for salvaged PR #17061
36131c5 
- Remove dead _lmstudio_loaded_context attribute from run_agent.py (set
  but never read — the loaded context is pushed to context_compressor.update_model
  which is the actual consumer)
- Cache empty reasoning options with 60s TTL to avoid per-turn HTTP probe
  for non-reasoning LM Studio models. Non-empty results cached permanently.
- Extract _lmstudio_server_root(), _lmstudio_request_headers(), and
  _lmstudio_fetch_raw_models() shared helpers in models.py — eliminates
  URL-strip + auth-header + HTTP-call duplication across probe_lmstudio_models,
  ensure_lmstudio_model_loaded, and lmstudio_model_reasoning_options
- Revert runtime_provider.py base_url precedence change: preserve the
  established contract (saved config.base_url > env var > default) for all
  api_key providers
- Remove unnecessary config version bump 22→23
- Fix TUI test: relax target_model assertion to avoid module-cache flake
- AUTHOR_MAP: added rugved@lmstudio.ai → rugvedS07
kshitijk4poor added a commit that referenced this pull request Apr 28, 2026
@kshitijk4poor
fix: follow-up for salvaged PR #17061
ca8f414 
- Remove dead _lmstudio_loaded_context attribute from run_agent.py (set
  but never read — the loaded context is pushed to context_compressor.update_model
  which is the actual consumer)
- Cache empty reasoning options with 60s TTL to avoid per-turn HTTP probe
  for non-reasoning LM Studio models. Non-empty results cached permanently.
- Extract _lmstudio_server_root(), _lmstudio_request_headers(), and
  _lmstudio_fetch_raw_models() shared helpers in models.py — eliminates
  URL-strip + auth-header + HTTP-call duplication across probe_lmstudio_models,
  ensure_lmstudio_model_loaded, and lmstudio_model_reasoning_options
- Revert runtime_provider.py base_url precedence change: preserve the
  established contract (saved config.base_url > env var > default) for all
  api_key providers
- Remove unnecessary config version bump 22→23
- Fix TUI test: relax target_model assertion to avoid module-cache flake
- AUTHOR_MAP: added rugved@lmstudio.ai → rugvedS07
kshitijk4poor added a commit that referenced this pull request Apr 28, 2026
@kshitijk4poor
fix: follow-up for salvaged PR #17061
5d2f9b5 
- Remove dead _lmstudio_loaded_context attribute from run_agent.py (set
  but never read — the loaded context is pushed to context_compressor.update_model
  which is the actual consumer)
- Cache empty reasoning options with 60s TTL to avoid per-turn HTTP probe
  for non-reasoning LM Studio models. Non-empty results cached permanently.
- Extract _lmstudio_server_root(), _lmstudio_request_headers(), and
  _lmstudio_fetch_raw_models() shared helpers in models.py — eliminates
  URL-strip + auth-header + HTTP-call duplication across probe_lmstudio_models,
  ensure_lmstudio_model_loaded, and lmstudio_model_reasoning_options
- Revert runtime_provider.py base_url precedence change: preserve the
  established contract (saved config.base_url > env var > default) for all
  api_key providers
- Remove unnecessary config version bump 22→23
- Fix TUI test: relax target_model assertion to avoid module-cache flake
- AUTHOR_MAP: added rugved@lmstudio.ai → rugvedS07
@kshitijk4poor

kshitijk4poor commented Apr 28, 2026

Copy link
Copy Markdown
Collaborator

Merged via PR #17102. Your commits were cherry-picked onto current main with your authorship preserved in git log. Follow-up fixes addressed self-review findings (dead code removal, reasoning cache TTL, DRY shared helpers, removed unnecessary config migration and scoped-out runtime_provider precedence change). Thank you for the excellent LM Studio integration — the JIT model loading and reasoning negotiation are features no competing agent offers!

ulasbilgen pushed a commit to ulasbilgen/hermes-adhd-agent that referenced this pull request May 1, 2026
@kshitijk4poor
fix: follow-up for salvaged PR NousResearch#17061
dbc7e15 
- Remove dead _lmstudio_loaded_context attribute from run_agent.py (set
  but never read — the loaded context is pushed to context_compressor.update_model
  which is the actual consumer)
- Cache empty reasoning options with 60s TTL to avoid per-turn HTTP probe
  for non-reasoning LM Studio models. Non-empty results cached permanently.
- Extract _lmstudio_server_root(), _lmstudio_request_headers(), and
  _lmstudio_fetch_raw_models() shared helpers in models.py — eliminates
  URL-strip + auth-header + HTTP-call duplication across probe_lmstudio_models,
  ensure_lmstudio_model_loaded, and lmstudio_model_reasoning_options
- Revert runtime_provider.py base_url precedence change: preserve the
  established contract (saved config.base_url > env var > default) for all
  api_key providers
- Remove unnecessary config version bump 22→23
- Fix TUI test: relax target_model assertion to avoid module-cache flake
- AUTHOR_MAP: added rugved@lmstudio.ai → rugvedS07
@subinium subinium mentioned this pull request May 11, 2026
Open
4 tasks
02356abc pushed a commit to 02356abc/hermes-agent that referenced this pull request May 14, 2026
@kshitijk4poor
fix: follow-up for salvaged PR NousResearch#17061
800b469 
- Remove dead _lmstudio_loaded_context attribute from run_agent.py (set
  but never read — the loaded context is pushed to context_compressor.update_model
  which is the actual consumer)
- Cache empty reasoning options with 60s TTL to avoid per-turn HTTP probe
  for non-reasoning LM Studio models. Non-empty results cached permanently.
- Extract _lmstudio_server_root(), _lmstudio_request_headers(), and
  _lmstudio_fetch_raw_models() shared helpers in models.py — eliminates
  URL-strip + auth-header + HTTP-call duplication across probe_lmstudio_models,
  ensure_lmstudio_model_loaded, and lmstudio_model_reasoning_options
- Revert runtime_provider.py base_url precedence change: preserve the
  established contract (saved config.base_url > env var > default) for all
  api_key providers
- Remove unnecessary config version bump 22→23
- Fix TUI test: relax target_model assertion to avoid module-cache flake
- AUTHOR_MAP: added rugved@lmstudio.ai → rugvedS07
dannyJ848 pushed a commit to dannyJ848/hermes-agent that referenced this pull request May 17, 2026
@kshitijk4poor
fix: follow-up for salvaged PR NousResearch#17061
92654f5 
- Remove dead _lmstudio_loaded_context attribute from run_agent.py (set
  but never read — the loaded context is pushed to context_compressor.update_model
  which is the actual consumer)
- Cache empty reasoning options with 60s TTL to avoid per-turn HTTP probe
  for non-reasoning LM Studio models. Non-empty results cached permanently.
- Extract _lmstudio_server_root(), _lmstudio_request_headers(), and
  _lmstudio_fetch_raw_models() shared helpers in models.py — eliminates
  URL-strip + auth-header + HTTP-call duplication across probe_lmstudio_models,
  ensure_lmstudio_model_loaded, and lmstudio_model_reasoning_options
- Revert runtime_provider.py base_url precedence change: preserve the
  established contract (saved config.base_url > env var > default) for all
  api_key providers
- Remove unnecessary config version bump 22→23
- Fix TUI test: relax target_model assertion to avoid module-cache flake
- AUTHOR_MAP: added rugved@lmstudio.ai → rugvedS07
gweeteve pushed a commit to gweeteve/hermes-agent that referenced this pull request Jun 2, 2026
@kshitijk4poor
fix: follow-up for salvaged PR NousResearch#17061
2477a80 
- Remove dead _lmstudio_loaded_context attribute from run_agent.py (set
  but never read — the loaded context is pushed to context_compressor.update_model
  which is the actual consumer)
- Cache empty reasoning options with 60s TTL to avoid per-turn HTTP probe
  for non-reasoning LM Studio models. Non-empty results cached permanently.
- Extract _lmstudio_server_root(), _lmstudio_request_headers(), and
  _lmstudio_fetch_raw_models() shared helpers in models.py — eliminates
  URL-strip + auth-header + HTTP-call duplication across probe_lmstudio_models,
  ensure_lmstudio_model_loaded, and lmstudio_model_reasoning_options
- Revert runtime_provider.py base_url precedence change: preserve the
  established contract (saved config.base_url > env var > default) for all
  api_key providers
- Remove unnecessary config version bump 22→23
- Fix TUI test: relax target_model assertion to avoid module-cache flake
- AUTHOR_MAP: added rugved@lmstudio.ai → rugvedS07
Egavasyug pushed a commit to Egavasyug/hermes-agent that referenced this pull request Jun 10, 2026
@kshitijk4poor
fix: follow-up for salvaged PR NousResearch#17061
3958007 
- Remove dead _lmstudio_loaded_context attribute from run_agent.py (set
  but never read — the loaded context is pushed to context_compressor.update_model
  which is the actual consumer)
- Cache empty reasoning options with 60s TTL to avoid per-turn HTTP probe
  for non-reasoning LM Studio models. Non-empty results cached permanently.
- Extract _lmstudio_server_root(), _lmstudio_request_headers(), and
  _lmstudio_fetch_raw_models() shared helpers in models.py — eliminates
  URL-strip + auth-header + HTTP-call duplication across probe_lmstudio_models,
  ensure_lmstudio_model_loaded, and lmstudio_model_reasoning_options
- Revert runtime_provider.py base_url precedence change: preserve the
  established contract (saved config.base_url > env var > default) for all
  api_key providers
- Remove unnecessary config version bump 22→23
- Fix TUI test: relax target_model assertion to avoid module-cache flake
- AUTHOR_MAP: added rugved@lmstudio.ai → rugvedS07
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

comp/agent Core agent loop, run_agent.py, prompt builder comp/cli CLI entry point, hermes_cli/, setup wizard comp/gateway Gateway runner, session dispatch, delivery comp/tui Terminal UI (ui-tui/ + tui_gateway/) P3 Low — cosmetic, nice to have type/feature New feature or request

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@rugvedS07 @kshitijk4poor @alt-glitch