Scott.Idem/Cortex-Inara
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: reasoning token budget + session name in header

Browse Source 
- reasoning_budget_tokens: optional int field on local_openai models;
  when set, injects {"reasoning": {"budget_tokens": N}} via extra_body
  into every OpenRouter API call (both tool-loop and confirmation-gate
  rounds). Field exposed in the model edit form in Settings.

- session name moved from standalone full-row div between #messages
  and #input-area into the persona-switcher block in the header, as a
  third dim line under "Cortex · Local". Collapses when empty via
  :empty CSS. No JS changes required.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Scott Idem
2026-05-09 21:35:23 -04:00
parent 85792a7bcf
commit a66c5a7f84
5 changed files with 80 additions and 57 deletions
Download Patch File Download Diff File Expand all files Collapse all files

39
cortex/model_registry.py
View File

@@ -664,7 +664,8 @@ def save_model(username: str, model_id: str | None, host_id: str,
label: str, model_name: str, context_k: int = 0, label: str, model_name: str, context_k: int = 0,
tags: list[str] | None = None, tags: list[str] | None = None,
max_rounds: int | None = None, max_rounds: int | None = None,
tools: bool = True) -> str: tools: bool = True,
reasoning_budget_tokens: int | None = None) -> str:
"""Create or update a local_openai model entry. Returns the model ID.""" """Create or update a local_openai model entry. Returns the model ID."""
data = _load(username) data = _load(username)
tags = tags or [] tags = tags or []
@@ -672,29 +673,31 @@ def save_model(username: str, model_id: str | None, host_id: str,
if model_id: if model_id:
for m in data["models"]: for m in data["models"]:
if m["id"] == model_id: if m["id"] == model_id:
m["host_id"] = host_id m["host_id"] = host_id
m["label"] = label.strip() or model_name.strip() m["label"] = label.strip() or model_name.strip()
m["model_name"] = model_name.strip() m["model_name"] = model_name.strip()
m["context_k"] = context_k m["context_k"] = context_k
m["max_rounds"] = max_rounds m["max_rounds"] = max_rounds
m["tools"] = tools m["tools"] = tools
m["tags"] = tags m["tags"] = tags
m["reasoning_budget_tokens"] = reasoning_budget_tokens
_save(username, data) _save(username, data)
return model_id return model_id
model_id = None model_id = None
model_id = secrets.token_hex(4) model_id = secrets.token_hex(4)
data["models"].append({ data["models"].append({
"id": model_id, "id": model_id,
"type": "local_openai", "type": "local_openai",
"label": label.strip() or model_name.strip(), "label": label.strip() or model_name.strip(),
"model_name": model_name.strip(), "model_name": model_name.strip(),
"provider": "local", "provider": "local",
"host_id": host_id, "host_id": host_id,
"context_k": context_k, "context_k": context_k,
"max_rounds": max_rounds, "max_rounds": max_rounds,
"tools": tools, "tools": tools,
"tags": tags, "tags": tags,
"reasoning_budget_tokens": reasoning_budget_tokens,
}) })
_save(username, data) _save(username, data)
return model_id return model_id

5
cortex/openai_orchestrator.py
View File

@@ -287,6 +287,9 @@ async def _run_from_messages(
if active_tools: if active_tools:
call_kwargs["tools"] = active_tools call_kwargs["tools"] = active_tools
call_kwargs["tool_choice"] = "auto" call_kwargs["tool_choice"] = "auto"
reasoning_budget = (model_cfg or {}).get("reasoning_budget_tokens")
if reasoning_budget:
call_kwargs["extra_body"] = {"reasoning": {"budget_tokens": reasoning_budget}}
response = await _chat_with_retry(client, **call_kwargs) response = await _chat_with_retry(client, **call_kwargs)
choice = response.choices[0] choice = response.choices[0]
@@ -346,6 +349,8 @@ async def _run_from_messages(
conf_call: dict = {"model": model_name, "messages": messages, "tool_choice": "none"} conf_call: dict = {"model": model_name, "messages": messages, "tool_choice": "none"}
if active_tools: if active_tools:
conf_call["tools"] = active_tools conf_call["tools"] = active_tools
if reasoning_budget:
conf_call["extra_body"] = {"reasoning": {"budget_tokens": reasoning_budget}}
conf_resp = await _chat_with_retry(client, **conf_call) conf_resp = await _chat_with_retry(client, **conf_call)
final_response = conf_resp.choices[0].message.content or ( final_response = conf_resp.choices[0].message.content or (
"This action requires your explicit confirmation before it can proceed." "This action requires your explicit confirmation before it can proceed."

82
cortex/routers/local_llm.py
View File

@@ -204,12 +204,13 @@ def _render(username: str, success: str = "", error: str = "") -> str:
else: else:
extra_fields = '<input type="hidden" name="credential_id" value="cli">' extra_fields = '<input type="hidden" name="credential_id" value="cli">'
cur_label = m.get("label", "") cur_label = m.get("label", "")
cur_model_name = m.get("model_name", "") cur_model_name = m.get("model_name", "")
cur_ctx = m.get("context_k", 0) or 0 cur_ctx = m.get("context_k", 0) or 0
cur_max_rounds = m.get("max_rounds") or 0 cur_max_rounds = m.get("max_rounds") or 0
cur_tools = m.get("tools", True) cur_tools = m.get("tools", True)
cur_tags = ", ".join(m.get("tags") or []) cur_tags = ", ".join(m.get("tags") or [])
cur_reasoning_budget = m.get("reasoning_budget_tokens") or 0
model_rows += f''' model_rows += f'''
<div class="model-row" id="model-{m["id"]}"> <div class="model-row" id="model-{m["id"]}">
@@ -256,6 +257,11 @@ def _render(username: str, success: str = "", error: str = "") -> str:
<input type="number" name="max_rounds" value="{cur_max_rounds}" min="0" <input type="number" name="max_rounds" value="{cur_max_rounds}" min="0"
title="Per-model tool loop cap. 0 = use the global default (orchestrator_max_rounds)."> title="Per-model tool loop cap. 0 = use the global default (orchestrator_max_rounds).">
</div> </div>
<div class="field" style="flex:0 0 auto">
<label title="OpenRouter reasoning budget in tokens. 0 = no reasoning override (model default). Injects reasoning.budget_tokens into the API call.">Reasoning tokens</label>
<input type="number" name="reasoning_budget_tokens" value="{cur_reasoning_budget}" min="0"
title="OpenRouter reasoning budget in tokens. 0 = disabled. E.g. 2048 for light thinking, 8192 for deep reasoning.">
</div>
<div class="field" style="flex:0 0 auto"> <div class="field" style="flex:0 0 auto">
<label title="Whether this model supports tool calling. If not supported, requests skip the tool loop entirely.">Tool calling</label> <label title="Whether this model supports tool calling. If not supported, requests skip the tool loop entirely.">Tool calling</label>
<select name="tools" <select name="tools"
@@ -459,20 +465,21 @@ async def remove_host(request: Request, host_id: str):
@router.post("/settings/local/models/add", include_in_schema=False) @router.post("/settings/local/models/add", include_in_schema=False)
async def add_model( async def add_model(
request: Request, request: Request,
provider: str = Form("local"), provider: str = Form("local"),
label: str = Form(""), label: str = Form(""),
context_k: int = Form(0), context_k: int = Form(0),
max_rounds: int = Form(0), max_rounds: int = Form(0),
tools: int = Form(1), tools: int = Form(1),
tags: str = Form(""), tags: str = Form(""),
reasoning_budget_tokens: int = Form(0),
# local-only fields # local-only fields
host_id: str = Form(""), host_id: str = Form(""),
model_name: str = Form(""), model_name: str = Form(""),
# cloud-only fields # cloud-only fields
cloud_model_name: str = Form(""), cloud_model_name: str = Form(""),
account_id: str = Form(""), account_id: str = Form(""),
credential_id: str = Form("cli"), credential_id: str = Form("cli"),
): ):
username = _get_user(request) username = _get_user(request)
if not username: if not username:
@@ -481,6 +488,7 @@ async def add_model(
tag_list = [t.strip() for t in tags.split(",") if t.strip()] tag_list = [t.strip() for t in tags.split(",") if t.strip()]
max_rounds_ = max_rounds or None max_rounds_ = max_rounds or None
tools_bool = tools != 0 tools_bool = tools != 0
reasoning_budget_ = reasoning_budget_tokens or None
if provider == "local": if provider == "local":
if not model_name.strip(): if not model_name.strip():
@@ -488,7 +496,8 @@ async def add_model(
if not host_id.strip(): if not host_id.strip():
return HTMLResponse(_render(username, error="Select a host.")) return HTMLResponse(_render(username, error="Select a host."))
reg.save_model(username, None, host_id, label, model_name, context_k, tag_list, reg.save_model(username, None, host_id, label, model_name, context_k, tag_list,
max_rounds=max_rounds_, tools=tools_bool) max_rounds=max_rounds_, tools=tools_bool,
reasoning_budget_tokens=reasoning_budget_)
display = label or model_name display = label or model_name
elif provider in ("google", "anthropic"): elif provider in ("google", "anthropic"):
@@ -513,32 +522,35 @@ async def add_model(
@router.post("/settings/local/models/{model_id}/edit", include_in_schema=False) @router.post("/settings/local/models/{model_id}/edit", include_in_schema=False)
async def edit_model( async def edit_model(
request: Request, request: Request,
model_id: str, model_id: str,
mtype: str = Form(""), mtype: str = Form(""),
label: str = Form(""), label: str = Form(""),
model_name: str = Form(""), model_name: str = Form(""),
context_k: int = Form(0), context_k: int = Form(0),
max_rounds: int = Form(0), max_rounds: int = Form(0),
tools: int = Form(1), tools: int = Form(1),
tags: str = Form(""), tags: str = Form(""),
host_id: str = Form(""), reasoning_budget_tokens: int = Form(0),
account_id: str = Form(""), host_id: str = Form(""),
credential_id: str = Form("cli"), account_id: str = Form(""),
credential_id: str = Form("cli"),
): ):
username = _get_user(request) username = _get_user(request)
if not username: if not username:
return RedirectResponse("/login", status_code=302) return RedirectResponse("/login", status_code=302)
if not model_name.strip(): if not model_name.strip():
return HTMLResponse(_render(username, error="Model name is required.")) return HTMLResponse(_render(username, error="Model name is required."))
tag_list = [t.strip() for t in tags.split(",") if t.strip()] tag_list = [t.strip() for t in tags.split(",") if t.strip()]
max_rounds_ = max_rounds or None max_rounds_ = max_rounds or None
tools_bool = tools != 0 tools_bool = tools != 0
reasoning_budget_ = reasoning_budget_tokens or None
if mtype == "local_openai": if mtype == "local_openai":
if not host_id.strip(): if not host_id.strip():
return HTMLResponse(_render(username, error="Select a host for this model.")) return HTMLResponse(_render(username, error="Select a host for this model."))
reg.save_model(username, model_id, host_id, label, model_name, context_k, tag_list, reg.save_model(username, model_id, host_id, label, model_name, context_k, tag_list,
max_rounds=max_rounds_, tools=tools_bool) max_rounds=max_rounds_, tools=tools_bool,
reasoning_budget_tokens=reasoning_budget_)
elif mtype == "gemini_api": elif mtype == "gemini_api":
reg.save_cloud_model(username, model_id, "google", model_name, label, reg.save_cloud_model(username, model_id, "google", model_name, label,
account_id=account_id or None, context_k=context_k, tags=tag_list, account_id=account_id or None, context_k=context_k, tags=tag_list,

2
cortex/static/index.html
View File

@@ -41,6 +41,7 @@
<div class="persona-switcher" id="persona-switcher"> <div class="persona-switcher" id="persona-switcher">
<div class="name" id="persona-name">Inara</div> <div class="name" id="persona-name">Inara</div>
<div class="subtitle">Cortex · Local</div> <div class="subtitle">Cortex · Local</div>
<div id="session-id"></div>
<div class="persona-dropdown" id="persona-dropdown"></div> <div class="persona-dropdown" id="persona-dropdown"></div>
</div> </div>
@@ -164,7 +165,6 @@
</div> </div>
<div id="messages"></div> <div id="messages"></div>
<div id="session-id"></div>
<div id="input-area"> <div id="input-area">
<!-- Mode select — compact dropdown, opens upward, MRU sorted --> <!-- Mode select — compact dropdown, opens upward, MRU sorted -->

9
cortex/static/style.css
View File

@@ -897,11 +897,14 @@
#stop:hover { background: #5c1a1a; } #stop:hover { background: #5c1a1a; }
#session-id { #session-id {
font-size: 0.7rem; font-size: 0.68rem;
color: var(--border); color: var(--border);
padding: 0 20px 6px; white-space: nowrap;
background: var(--surface); overflow: hidden;
text-overflow: ellipsis;
max-width: 220px;
} }
#session-id:empty { display: none; }
/* ── Message wrappers (edit/delete controls) ──────────────── */ /* ── Message wrappers (edit/delete controls) ──────────────── */
.msg-wrapper { .msg-wrapper {