Configuration
Configure AI Council: environment variables, AI providers, and model selection for every agent on the bench.
Environment Variables
Configure AI Council's behavior through environment variables.
| Variable | Required | Default | Description |
|---|---|---|---|
AI_COUNCIL_LICENSE_KEY |
Yes | - | Your license key from the store |
OPENAI_API_KEY |
No* | - | OpenAI API key for GPT models |
GEMINI_API_KEY |
No* | - | Google Gemini API key |
AI_COUNCIL_MODEL |
No | gpt-5.4-mini |
OpenAI model to use |
AI_COUNCIL_GEMINI_MODEL |
No | gemini-3.6-flash |
Gemini model to use |
AI_COUNCIL_SUGGESTION_CLASSIFY_MAX |
No | 20 |
Max suggestion-classify calls per window in chat |
AI_COUNCIL_SUGGESTION_CLASSIFY_WINDOW_SECONDS |
No | 60 |
Window in seconds for the above |
AI_COUNCIL_PRE_CONTEXT |
No | false |
Enable pre-review AST + embedding context (true/1) |
AI_COUNCIL_PRE_CONTEXT_TOP_K |
No | 5 |
Number of top code units to inject |
AI_COUNCIL_PRE_CONTEXT_MAX_CHARS |
No | 1500 |
Max chars per code snippet in pre-context |
AI_COUNCIL_PRE_CONTEXT_TIMEOUT |
No | 15000 |
Timeout in ms for pre-context pipeline |
AI_COUNCIL_FULL_FILE_MAX_BYTES |
No | impl default | Max bytes for a single full-file include |
AI_COUNCIL_FULL_FILE_BUDGET_BYTES |
No | impl default | Total byte budget across full-file includes |
AI_COUNCIL_MAX_FILES |
No | impl default | Max files reviewed from the selected diff source |
AI_COUNCIL_GRAPH |
No | auto | 0/false disables; 1/true forces attempt (default: auto when discovery finds a source) |
AI_COUNCIL_GRAPH_MAX_CHARS |
No | 12288 |
Max chars in the graph context block |
NEO4J_URI / NEO4J_URL |
No | - | Neo4j/Aura Bolt URI |
NEO4J_USERNAME / NEO4J_USER |
No | - | Neo4j username |
NEO4J_PASSWORD |
No | - | Neo4j password |
NEO4J_DATABASE |
No | - | Optional Neo4j database |
MEMGRAPH_URI / MEMGRAPH_URL |
No | - | Memgraph Bolt URI |
MEMGRAPH_USERNAME / MEMGRAPH_USER |
No | - | Memgraph username |
MEMGRAPH_PASSWORD |
No | - | Memgraph password |
MEMGRAPH_DATABASE |
No | - | Optional Memgraph database |
AI_COUNCIL_GRAPH_HTTP_URL |
No | - | HTTP graph query endpoint |
AI_COUNCIL_GRAPH_HTTP_HEADERS_JSON |
No | - | Optional JSON object of HTTP headers |
* At least one of OPENAI_API_KEY or GEMINI_API_KEY is required.
AI Providers
AI Council supports multiple AI providers for diverse perspectives.
OpenAI (Primary)
Used for most agents including Senior Developer, Security, Performance, Architect, and Pragmatist.
export OPENAI_API_KEY="sk-..."Google Gemini
Powers the Gemini Structural Thinker agent, providing alternative reasoning and challenging assumptions.
export GEMINI_API_KEY="..."
Best Practice
Configure both providers for the most comprehensive reviews. The diversity of AI models leads to better coverage and catches more edge cases.
Model Selection
Customize which AI models power your council.
OpenAI Models
| Model | Speed | Cost | Quality |
|---|---|---|---|
gpt-5.4-mini |
Fast | Low | Good (default) |
gpt-5.4-nano |
Fastest | Lowest | Good |
gpt-5.4 |
Medium | High | Excellent |
Gemini Models
| Model | Speed | Cost | Quality |
|---|---|---|---|
gemini-3.6-flash |
Fast | Low | Good (default) |
gemini-3.5-flash |
Fast | Low | Good |
Need Help?
Can't find what you're looking for? Here are some resources: