Google Interactions API Coverage Report

Spec version: v1beta

This page provides a detailed breakdown of Llama Stack's coverage of the Google Interactions API specification. The coverage score increases as missing features are implemented.

Auto-generated

This documentation is auto-generated from the Google Interactions API specification comparison.

Summary

Metric	Value
Overall Coverage Score	32.9%
Total Items Implemented	27/82
Total Missing	55

Section Scores

Sections are sorted by coverage score (lowest first, needing most attention).

Section	Score	Implemented	Total	Missing
Tool Types	0.0%	0	9	9
Content Types	5.0%	1	20	19
Endpoints	25.0%	1	4	3
GenerationConfig	30.0%	3	10	7
Usage	30.0%	3	10	7
Request Properties	46.2%	6	13	7
Streaming Events	71.4%	5	7	2
Response Properties	88.9%	8	9	1

Detailed Breakdown

Below is a detailed breakdown of what is implemented and what is missing for each section.

Tool Types

Score: 0.0% · Implemented: 0/9

Missing (9)

• CodeExecution
• ComputerUse
• FileSearch
• Function
• GoogleMaps
• GoogleSearch
• McpServer
• Retrieval
• UrlContext

Content Types

Score: 5.0% · Implemented: 1/20

Implemented (1)

• TextContent

Missing (19)

• AudioContent
• CodeExecutionCallContent
• CodeExecutionResultContent
• DocumentContent
• FileSearchCallContent
• FileSearchResultContent
• FunctionCallContent
• FunctionResultContent
• GoogleMapsCallContent
• GoogleMapsResultContent
• GoogleSearchCallContent
• GoogleSearchResultContent
• ImageContent
• McpServerToolCallContent
• McpServerToolResultContent
• ThoughtContent
• UrlContextCallContent
• UrlContextResultContent
• VideoContent

Endpoints

Score: 25.0% · Implemented: 1/4

Implemented (1)

• POST /{api_version}/interactions

Missing (3)

• DELETE /{api_version}/interactions/{id}
• GET /{api_version}/interactions/{id}
• POST /{api_version}/interactions/{id}/cancel

GenerationConfig

Score: 30.0% · Implemented: 3/10

Implemented (3)

• max_output_tokens
• temperature
• top_p

Missing (7)

• image_config
• seed
• speech_config
• stop_sequences
• thinking_level
• thinking_summaries
• tool_choice

Extra in Llama Stack (1)

These properties are in the Llama Stack implementation but not in the Google spec:

• top_k

Usage

Score: 30.0% · Implemented: 3/10

Implemented (3)

• total_input_tokens
• total_output_tokens
• total_tokens

Missing (7)

• cached_tokens_by_modality
• input_tokens_by_modality
• output_tokens_by_modality
• tool_use_tokens_by_modality
• total_cached_tokens
• total_thought_tokens
• total_tool_use_tokens

Request Properties

Score: 46.2% · Implemented: 6/13

Implemented (6)

• generation_config
• input
• model
• response_modalities
• stream
• system_instruction

Missing (7)

• background
• previous_interaction_id
• response_format
• response_mime_type
• service_tier
• store
• tools

Streaming Events

Score: 71.4% · Implemented: 5/7

Implemented (5)

• ContentDelta
• ContentStart
• ContentStop
• InteractionCompleteEvent
• InteractionStartEvent

Missing (2)

• ErrorEvent
• InteractionStatusUpdate

Response Properties

Score: 88.9% · Implemented: 8/9

Implemented (8)

• created
• id
• model
• outputs
• role
• status
• updated
• usage

Missing (1)

• agent

Extra in Llama Stack (1)

These properties are in the Llama Stack implementation but not in the Google spec:

• object

How to Improve Coverage

To improve coverage scores:

1. Add Missing Properties: Implement missing fields in request/response models in src/llama_stack_api/interactions/models.py
2. Add Content Types: Support additional content types beyond text (images, audio, function calls, etc.)
3. Add Tool Support: Implement tool declarations (Function, GoogleSearch, CodeExecution, etc.)
4. Add Missing Endpoints: Implement GET, DELETE, and Cancel endpoints

Run the coverage analyzer to check your progress:

python scripts/google_interactions_coverage.py --update --generate-docs