Your task is to analyze a sequence of images taken in chronological order from a security camera.
## Normal Activity Patterns for This Property
{activity_context_prompt}
## Task Instructions
Describe the scene based on observable actions and movements, evaluate the activity against the Activity Indicators above, and assign a potential_threat_level (0, 1, or 2) by applying the threat level indicators consistently.
- **Treat "Objects in Scene" as the list of tracked subjects to describe.** Do not introduce additional people or vehicles that are not present in this list. You may freely reference other items, surfaces, and environmental details visible in the frames when describing what the listed subjects are doing.
- **Describe the most likely activity from visible cues across the sequence** — the subject's path, what they are carrying, and what they interact with. Avoid asserting completed outcomes you do not observe; describe in-progress actions rather than results.
- Describe what you observe: actions, movements, interactions with objects and the environment. Include any observable environmental changes (e.g., lighting changes triggered by activity).
- Note visible details such as clothing, items being carried or placed, tools or equipment present, and how they interact with the property or objects.
- Consider the full sequence chronologically: what happens from start to finish, how duration and actions relate to the location and objects involved.
- **Use the actual timestamp provided in "Activity started at"** below for time of day context—do not infer time from image brightness or darkness. Unusual hours (late night/early morning) should increase suspicion when the observable behavior itself appears questionable. However, recognize that some legitimate activities can occur at any hour.
- **Consider duration as a primary factor**: Apply the duration thresholds defined in the activity patterns above. Brief sequences during normal hours with apparent purpose typically indicate normal activity unless explicit suspicious actions are visible.
- **Weigh all evidence holistically**: Match the activity against the normal and suspicious patterns defined above, then evaluate based on the complete context (zone, objects, time, actions, duration). Apply the threat level indicators consistently. Use your judgment for edge cases.
## Response Field Guidelines
Respond with a JSON object matching the provided schema. Field-specific guidance:
- `observations`: Include the very start of the activity — for example, a vehicle entering the frame or pulling into the driveway — even if it lasts only a few frames and the rest of the clip is dominated by a longer activity. Include each arrival, departure, object handled, and notable change in position or state. Each item is a single concrete fact written as a complete sentence.
- `scene`: Describe how the sequence begins, then the progression of events — all significant movements and actions in order. For example, if a vehicle arrives and then a person exits, describe both sequentially. For named subjects (those with a `←` separator in "Objects in Scene"), always use their name — do not replace them with generic terms. For unnamed objects (e.g., "person", "car"), refer to them naturally with articles (e.g., "a person", "the car"). Your description should align with and support the threat level you assign.
- `title`: Name the primary activity across the observations, together with the location. An activity is what is being done with objects, tools, or surfaces; locomotion through the scene qualifies as the activity only when no other interaction is observed. For named subjects, always use their name. For unnamed objects, refer to them naturally with articles.
- `shortSummary`: Briefly summarize the primary activity across the observations.
- `potential_threat_level`: Must be consistent with your scene description and the activity patterns above.
- Activity started at {review_data["start"]} and lasted {review_data["duration"]} seconds
- Zones involved: {", ".join(review_data["zones"])ifreview_data["zones"]else"None"}
## Objects in Scene
Each line represents a detection state, not necessarily unique individuals. The `←` symbol separates a recognized subject's name from their object type — use only the name (before the `←`) in your response, not the type after it. The same subject may appear across multiple lines if detected multiple times.
**Note: Unidentified objects (without names) are NOT indicators of suspicious activity—they simply mean the system hasn't identified that object.**
- "context": array of related events from other cameras that occurred during overlapping time periods
**Note: Use the "scene" field for event descriptions in the report. Ignore any "shortSummary" field if present.**
Report Structure - Use this EXACT format:
# Security Summary - {time_range}
## Overview
[Write 1-2 sentences summarizing the overall activity pattern during this period.]
---
## Timeline
[Group events by time periods (e.g., "Morning (6:00 AM - 12:00 PM)", "Afternoon (12:00 PM - 5:00 PM)", "Evening (5:00 PM - 9:00 PM)", "Night (9:00 PM - 6:00 AM)"). Use appropriate time blocks based on when events occurred.]
- [Event title]: [Clear description incorporating contextual information from the "context" array]
- Context: [If context array has items, mention them here, e.g., "Delivery truck present on Front Driveway Cam (HH:MM AM/PM)"]
- Assessment: [Brief assessment incorporating context - if context explains the event, note it here]
[Repeat for each event in chronological order within the time block]
---
## Summary
[One sentence summarizing the period. If all events are normal/explained: "Routine activity observed." If review needed: "Some activity requires review but no security concerns." If security concerns: "Security concerns requiring immediate attention."]
Guidelines:
- List ALL events in chronological order, grouped by time blocks
- Integrate contextual information naturally - use the "context" array to enrich each event's description
- If context explains the event (e.g., delivery truck explains person at door), describe it accordingly (e.g., "delivery person" not "unidentified person")
- Be concise but informative - focus on what happened and what it means
- If contextual information makes an event clearly normal, reflect that in your assessment
- Only create time blocks that have events - don't create empty sections
"""
prompt+="\n\nEvents:\n"
foreventinevents:
prompt+=f"\n{event}\n"
ifpreferred_language:
prompt+=f"\nProvide your answer in {preferred_language}"
returnprompt
defbuild_object_description_prompt(
camera_config:CameraConfig,
event:Event,
)->str:
"""Build the prompt for a per-object description.
Pulls the per-label override from `objects.genai.object_prompts`, falling
back to the camera default, and interpolates event fields.
Raises:
KeyError: if the user-defined prompt template references an unknown
+"\n\nWhen users refer to cameras or zones by their friendly name (e.g., 'Back Deck Camera', 'Front Walkway'), use the corresponding ID (e.g., 'back_deck_cam', 'front_walk') in tool calls. Tool results also identify zones by their ID, so when presenting cameras or zones back to the user, translate the ID to its friendly name."
speed_units_section=f"\n\nReport object speeds to the user in {speed_unit}."
semantic_search_section=""
ifsemantic_search_enabled:
semantic_search_section=(
"\n\nWhen routing a search_objects call, pick filters by the shape of the user's request:\n"
"- Generic class ('show me all cars today'): set `label` only.\n"
"- Specific named entity — a known person ('John'), delivery company ('Amazon'), animal species/breed ('blue jay', 'cardinal', 'golden retriever'), or license plate: set `sub_label` only and leave `label` unset.\n"
"- Physical characteristic, appearance, or activity that is NOT a discrete name ('find me people riding a lawn mower', 'someone in a red jacket', 'a person carrying a package'): set `semantic_query` with the descriptive phrase, optionally combined with `label` for the object class. Never put descriptive phrases in `sub_label`."
)
attribute_classification_section=""
ifattribute_classifications:
model_lines="\n".join(
f"- {m['name']}: applies to {', '.join(m['objects'])or'any object'}"
forminattribute_classifications
)
attribute_classification_section=(
"\n\nAttribute classification models are configured for the following object types:\n"
f"{model_lines}\n"
"When the user's request matches one of these classifications, set the search_objects `attribute` field to the matching label rather than using `semantic_query`. Reserve `semantic_query` for descriptive phrases that fall outside the configured attribute labels."
)
returnf"""You are a helpful assistant for Frigate, a security camera NVR system. You help users answer questions about their cameras, detected objects, and events.
Current server local date and time: {current_date_str} at {current_time_str}
Do not start your response with phrases like "I will check...", "Let me see...", or "Let me look...". Answer directly.
Always present times to the user in the server's local timezone. When tool results include start_time_local and end_time_local, use those exact strings when listing or describing detection times—do not convert or invent timestamps. Do not use UTC or ISO format with Z for the user-facing answer unless the tool result only provides Unix timestamps without local time fields.
When users ask about "today", "yesterday", "this week", etc., use the current date above as reference.
When searching for objects or events, use ISO 8601 format for dates (e.g., {current_date_str}T00:00:00Z for the start of today).
Always be accurate with time calculations based on the current date provided.
When a user refers to a specific object they have seen or describe with identifying details ("that green car", "the person in the red jacket", "a package left today"), prefer the find_similar_objects tool over search_objects. Use search_objects first only to locate the anchor event, then pass its id to find_similar_objects. For generic queries like "show me all cars today", keep using search_objects. If a user message begins with [attached_event:<id>], treat that event id as the anchor for any similarity or "tell me more" request in the same message and call find_similar_objects with that id.{semantic_search_section}{attribute_classification_section}{cameras_section}{speed_units_section}"""
