Big change

.gitignore

@@ -49,7 +49,7 @@ env/
 *.sqlite
 *.sqlite3
 data/config.json
-data/llm_models.json
+llm_models.json
 data/whisper-models/*
 !data/whisper-models/.gitkeep
 data/piper-voices/*

CHANGELOG.md

@@ -16,6 +16,29 @@ Work in progress is accumulated under `[Unreleased]`; on release, that section b
 
 ### Removed
 
+## 2026.6.16
+
+### Added
+
+- **Known questions** — mark theory or coding bank items as known during an interview (**I know this**) or on review pages; optionally exclude them when starting a new session; manage the list from **Known Questions** in the navigation bar
+
+### Changed
+
+- **Add model to catalog** — the catalog model id is generated automatically from the display name; removed the **Model id** field on `/config`
+- **UI** — refreshed dark theme with clearer hierarchy, IDE-style coding editor, terminal-style run output, and updated status badges on the dashboard
+
+### Fixed
+
+- **Theory then coding sessions** — fixed errors when advancing from theory to coding in a combined session
+- **Coding follow-ups** — explanation rounds now submit your typed explanation instead of the code in the editor
+- **Coding timers** — expired rounds score 0 and the session advances automatically
+- **Setup review** — known-questions option shows the correct hint for the checkbox
+- **Early session end** — partial theory/coding scores are kept when you end a session before finishing every task
+- **Theory answers** — more reliable submit flow for text and audio answers during AI evaluation
+- **Dashboard** — faster interview history on the home page
+
+### Removed
+
 ## 2026.6.12
 
 ### Added

README.md

@@ -18,6 +18,7 @@ A general chat assistant is flexible, but it does not run an **interview** for y
 | Interview flow | Free-form thread | Fixed session: theory Q&A and/or coding tasks, up to **2 AI follow-ups** per item, **1–5 scoring**, session summary |
 | Live coding practice | Paste code in chat | **Monaco editor**, **Run** against public tests, **Submit** for hidden tests + AI review (needs Judge0) |
 | Practice history | Scattered chats | **Dashboard** with past sessions; open **results** and per-section **review** pages after completion |
+| Skip what you already know | You repeat the same prompts | **Known questions** — mark bank items during practice; optionally exclude them when starting a new session |
 | Time pressure | None | Optional **per-round timer** on theory and coding (expired round → 0, move on) |
 | Voice practice | Depends on product | Offline **Whisper** dictation; optional **Piper** question audio; **audio answers** when your model supports it |
 | Where data lives | Vendor cloud | **Self-hosted**: SQLite + `data/` on your machine; use **Ollama**, vLLM, or any OpenAI-compatible API |
@@ -31,7 +32,10 @@ A general chat assistant is flexible, but it does not run an **interview** for y
 **Demo video** — full flow from setup to scored feedback
 
 <p align="center">
-  <img src="./assets/demo_cut.gif" alt="Demo video" width="900" />
+  <video src="./assets/demo-video.mp4" width="900" controls playsinline>
+    Your browser does not support the video tag.
+    <a href="./assets/demo-video.mp4">Download the demo video</a>.
+  </video>
 </p>
 
 **Dashboard** — recent sessions and quick start
@@ -82,6 +86,7 @@ Coding modes need a running [Judge0](https://github.com/judge0/judge0) instance
 - **Voice** — offline Whisper dictation; optional Piper TTS to read theory questions aloud
 - **Audio answers** — record a WAV theory answer when your model supports audio input and Whisper is ready
 - **Results hub** — after you finish, `/interview/{id}/results` shows overall evaluation and links to **theory** and **coding** review pages with full chat/code history
+- **Known questions** — mark theory or coding bank items as **I know this** during an interview or on review pages; optionally exclude them on **New interview** setup; manage the list at `/known-questions/manage`
 - **Dashboard** — recent sessions on the home page (completed sessions link to results)
 - **Setup** — model catalog on `/config`, interview locale, Whisper/Piper downloads from the UI
 - **Deployment** — Docker Compose on port 8000 with `./data` volume for config, DB, and models
@@ -131,7 +136,7 @@ On some Linux hosts Judge0 needs **cgroup v1** (`systemd.unified_cgroup_hierarch
 ### First-time flow
 
 1. **Configuration** (`/config`) — add one or more OpenAI-compatible models to the catalog, select an interview model, set interview locale; test connection, then save. Download Whisper (and optionally a Piper voice) from the same page if you want voice features.
-2. **New interview** (`/setup`) — pick a **session mode** (theory only, coding only, or combined). Choose tracks, levels, topics, how many questions/tasks, and optional per-round timers. Coding modes require Judge0 (see **Coding sessions** above).
+2. **New interview** (`/setup`) — pick a **session mode** (theory only, coding only, or combined). Choose tracks, levels, topics, how many questions/tasks, optional per-round timers, and whether to **exclude known questions**. Coding modes require Judge0 (see **Coding sessions** above).
 3. **Practice** (`/interview/{id}`) — answer theory questions in the chat (type, dictate, or record audio). On coding phases, use the editor: **Run** to check public tests, **Submit** when ready. Combined sessions switch panels automatically when a section ends (or use **Continue to Coding**). End the interview from the sidebar at any time.
 4. **Review** (`/interview/{id}/results`) — after completion, read the overall evaluation, then open **Theory** or **Coding** review for full conversation history, scores, and feedback.
 
@@ -160,7 +165,7 @@ Any **OpenAI-compatible** HTTP API works:
 
 On `/config`:
 
-- **Add model to catalog** — base URL, model name, optional API key; enable **Accepts audio input** only if the model supports multimodal audio (and download Whisper for transcription).
+- **Add model to catalog** — display name, base URL, model name, optional API key (a stable catalog id is generated automatically from the display name); enable **Accepts audio input** only if the model supports multimodal audio (and download Whisper for transcription).
 - **Interview model** — pick from the catalog, **Test Connection**, save.
 - **Locale** — language for AI feedback and speech (stored in `data/config.json`, gitignored).
 - **Whisper** — choose size (`small`, `medium`, `large`), download from the UI for dictation and audio answers.

alembic/versions/20260612_0010_answers_expected_points.py

@@ -0,0 +1,26 @@
+# Copyright 2026 GrillKit Contributors
+# SPDX-License-Identifier: Apache-2.0
+"""Add expected_points rubric snapshot column to answers."""
+
+from collections.abc import Sequence
+
+import sqlalchemy as sa
+
+from alembic import op
+
+revision: str = "20260612_0010"
+down_revision: str | None = "20260610_0009"
+branch_labels: str | Sequence[str] | None = None
+depends_on: str | Sequence[str] | None = None
+
+
+def upgrade() -> None:
+    """Store rubric bullets on each theory answer row."""
+    with op.batch_alter_table("answers") as batch_op:
+        batch_op.add_column(sa.Column("expected_points", sa.Text(), nullable=True))
+
+
+def downgrade() -> None:
+    """Remove expected_points from answers."""
+    with op.batch_alter_table("answers") as batch_op:
+        batch_op.drop_column("expected_points")

alembic/versions/20260615_0011_known_questions.py

@@ -0,0 +1,35 @@
+# Copyright 2026 GrillKit Contributors
+# SPDX-License-Identifier: Apache-2.0
+"""Add known_questions table for excluding marked questions from planning."""
+
+from collections.abc import Sequence
+
+import sqlalchemy as sa
+
+from alembic import op
+
+revision: str = "20260615_0011"
+down_revision: str | None = "20260612_0010"
+branch_labels: str | Sequence[str] | None = None
+depends_on: str | Sequence[str] | None = None
+
+
+def upgrade() -> None:
+    """Create known_questions table with composite primary key."""
+    op.create_table(
+        "known_questions",
+        sa.Column("branch", sa.Text(), nullable=False),
+        sa.Column("bank_item_id", sa.Text(), nullable=False),
+        sa.Column(
+            "created_at",
+            sa.DateTime(timezone=True),
+            server_default=sa.text("CURRENT_TIMESTAMP"),
+            nullable=False,
+        ),
+        sa.PrimaryKeyConstraint("branch", "bank_item_id"),
+    )
+
+
+def downgrade() -> None:
+    """Drop known_questions table."""
+    op.drop_table("known_questions")

app/ai/llm_models.py

@@ -2,14 +2,13 @@
 # SPDX-License-Identifier: Apache-2.0
 """LLM model catalog types."""
 
+from collections.abc import Collection
 from dataclasses import dataclass
 import re
 from typing import Final
 
 CUSTOM_PRESET_ID: Final[str] = "custom"
-MODEL_ID_PATTERN: Final[re.Pattern[str]] = re.compile(
-    r"^[a-z0-9](?:[a-z0-9-]*[a-z0-9])?$"
-)
+_FALLBACK_MODEL_ID: Final[str] = "model"
 
 
 @dataclass(frozen=True)
@@ -71,25 +70,39 @@ def normalize_model_id(model_id: str, catalog: LLMCatalog) -> str:
     return value
 
 
-def validate_new_model_id(model_id: str) -> str:
-    """Normalize and validate a user-supplied catalog model id.
+def slugify_model_id(text: str) -> str:
+    """Convert arbitrary text into a catalog-safe id slug.
+
+    Lowercases the text and replaces every run of non-alphanumeric characters
+    with a single hyphen, trimming hyphens from both ends.
 
     Args:
-        model_id: Proposed model id from the add-model form.
+        text: Source text, typically a model display name.
 
     Returns:
-        Normalized lowercase id.
+        A slug using lowercase letters, digits, and hyphens, or an empty
+        string when no usable characters remain.
+    """
+    return re.sub(r"[^a-z0-9]+", "-", text.strip().lower()).strip("-")
 
-    Raises:
-        ValueError: If the id is invalid or reserved.
+
+def generate_model_id(display_name: str, existing_ids: Collection[str]) -> str:
+    """Derive a unique catalog id from a model display name.
+
+    Args:
+        display_name: Human-readable model name from the add-model form.
+        existing_ids: Catalog ids already in use.
+
+    Returns:
+        A unique slug; falls back to ``model`` when the name has no usable
+        characters and appends a numeric suffix to avoid collisions.
     """
-    value = model_id.strip().lower()
-    if not value:
-        raise ValueError("Model id is required")
-    if value == CUSTOM_PRESET_ID:
-        raise ValueError(f"Model id '{CUSTOM_PRESET_ID}' is reserved")
-    if not MODEL_ID_PATTERN.fullmatch(value):
-        raise ValueError(
-            "Model id must use lowercase letters, digits, and hyphens only"
-        )
-    return value
+    base = slugify_model_id(display_name) or _FALLBACK_MODEL_ID
+    if base == CUSTOM_PRESET_ID:
+        base = f"{CUSTOM_PRESET_ID}-{_FALLBACK_MODEL_ID}"
+    candidate = base
+    suffix = 2
+    while candidate in existing_ids:
+        candidate = f"{base}-{suffix}"
+        suffix += 1
+    return candidate

app/coding/api/errors.py

@@ -12,7 +12,11 @@
     CodingTaskNotCurrentError,
     CodingTaskNotFoundError,
 )
-from app.interview.domain.exceptions import InterviewDomainError
+from app.interview.domain.exceptions import (
+    InterviewDomainError,
+    InterviewNotActiveError,
+    InterviewNotFoundError,
+)
 
 
 def coding_ws_error_payload(
@@ -42,14 +46,16 @@ def http_exception_from_coding_error(
     """
     if isinstance(
         exc,
-        CodingSectionNotFoundError | CodingTaskNotFoundError,
+        InterviewNotFoundError | CodingSectionNotFoundError | CodingTaskNotFoundError,
     ):
         return HTTPException(status_code=404, detail=str(exc))
     if isinstance(exc, CodingRunLimitExceededError):
         return HTTPException(status_code=429, detail=str(exc))
     if isinstance(
         exc,
-        CodingSectionNotActiveError | CodingTaskNotCurrentError,
+        InterviewNotActiveError
+        | CodingSectionNotActiveError
+        | CodingTaskNotCurrentError,
     ):
         return HTTPException(status_code=400, detail=str(exc))
     return HTTPException(status_code=400, detail=str(exc))

app/coding/api/routes.py

@@ -19,8 +19,11 @@
     run_attempt_to_response,
 )
 from app.coding.services.run_execution import CodingRunExecutionService
-from app.coding.services.state import CodingStateService
-from app.interview.api.deps import AIProviderDep
+from app.interview.api.deps import (
+    AIProviderDep,
+    CodingStateServiceDep,
+    CodingSubmissionServiceDep,
+)
 from app.interview.domain.exceptions import InterviewDomainError
 
 router = APIRouter(prefix="/interview", tags=["coding"])
@@ -74,7 +77,10 @@ async def coding_run(
 
 
 @router.get("/{interview_id}/coding/state")
-async def coding_state(interview_id: str) -> JSONResponse:
+async def coding_state(
+    interview_id: str,
+    service: CodingStateServiceDep,
+) -> JSONResponse:
     """Return coding session progress and Run history for the active task.
 
     Args:
@@ -87,7 +93,7 @@ async def coding_state(interview_id: str) -> JSONResponse:
         HTTPException: When the coding section does not exist.
     """
     try:
-        state = CodingStateService.get_state(interview_id)
+        state = service.get_state(interview_id)
     except CodingDomainError as exc:
         raise http_exception_from_coding_error(exc) from exc
     return JSONResponse(coding_state_to_dict(state))
@@ -98,6 +104,7 @@ async def coding_ws(
     websocket: WebSocket,
     interview_id: str,
     provider: AIProviderDep,
+    submission_service: CodingSubmissionServiceDep,
 ) -> None:
     """WebSocket endpoint for coding task submit and feedback.
 
@@ -118,6 +125,7 @@ async def coding_ws(
                 raw,
                 interview_id=interview_id,
                 provider=provider,
+                submission_service=submission_service,
             ):
                 if not await _safe_send_json(websocket, message):
                     break

app/coding/api/ws_session.py

@@ -26,15 +26,15 @@ async def iter_responses(
         *,
         interview_id: str,
         provider: AIProvider,
-        submission_service: type[CodingSubmissionService] = CodingSubmissionService,
+        submission_service: CodingSubmissionService,
     ) -> AsyncIterator[dict[str, Any]]:
         """Handle one client message and yield JSON payloads for the socket.
 
         Args:
             raw: Parsed client JSON message.
             interview_id: Interview session UUID.
             provider: AI provider for coding evaluation.
-            submission_service: Coding submission service class.
+            submission_service: Request-scoped coding submission service.
 
         Yields:
             WebSocket message dicts to send to the client.
@@ -50,6 +50,15 @@ async def iter_responses(
                 yield message
             return
 
+        if msg_type == "timeout":
+            async for message in CodingWebSocketService._handle_timeout(
+                raw,
+                interview_id=interview_id,
+                submission_service=submission_service,
+            ):
+                yield message
+            return
+
         yield {
             "type": "error",
             "message": f"Unknown message type: {msg_type}",
@@ -61,7 +70,7 @@ async def _handle_submit(
         *,
         interview_id: str,
         provider: AIProvider,
-        submission_service: type[CodingSubmissionService],
+        submission_service: CodingSubmissionService,
     ) -> AsyncIterator[dict[str, Any]]:
         task_id = str(raw.get("task_id", "")).strip()
         source_code = str(raw.get("source_code", ""))
@@ -88,3 +97,32 @@ async def _handle_submit(
                 "type": "error",
                 "message": ai_error_message_for_client(exc),
             }
+
+    @staticmethod
+    async def _handle_timeout(
+        raw: dict[str, Any],
+        *,
+        interview_id: str,
+        submission_service: CodingSubmissionService,
+    ) -> AsyncIterator[dict[str, Any]]:
+        task_id = str(raw.get("task_id") or raw.get("question_id") or "").strip()
+        round_num = raw.get("round")
+        if not task_id or round_num is None:
+            yield {"type": "error", "message": "Both task_id and round are required"}
+            return
+
+        try:
+            async for event in submission_service.stream_timeout_submission(
+                interview_id=interview_id,
+                task_id=task_id,
+                round_num=int(round_num),
+            ):
+                yield coding_event_to_message(event)
+        except (InterviewDomainError, CodingDomainError) as exc:
+            yield coding_ws_error_payload(exc)
+        except Exception as exc:
+            logger.exception("Coding timeout failed for interview %s", interview_id)
+            yield {
+                "type": "error",
+                "message": ai_error_message_for_client(exc),
+            }