Documentation
Configuring ESBMC-AI

Configuring ESBMC-AI

Info

You can find the default config here. You can start configuring ESBMC-AI after this.

The command esbmc-ai help-config will list all of the config fields along with help messages that explain what each one does.

Notice

The help-config command does not list the help message of addons currently, this is a limitation.

Configuration Sources and Precedence

ESBMC-AI accepts configuration from multiple sources with the following precedence (highest to lowest):

CLI arguments > Environment variables > .env file > TOML config file > Defaults

This means CLI arguments will override environment variables, which override values in the .env file, and so on.

Setting Nested Configuration Values

Nested configuration fields can be set via environment variables using double underscores (__):

# Set verifier.esbmc.path
export ESBMCAI_VERIFIER__ESBMC__PATH=/usr/bin/esbmc

# Set verifier.esbmc.timeout
export ESBMCAI_VERIFIER__ESBMC__TIMEOUT=30

All environment variables must use the ESBMCAI_ prefix.

Configuring AI Models

ESBMC-AI supports all LangChain-compatible LLM providers through the universal init_chat_model interface. Built-in support includes OpenAI, Anthropic, Ollama, and any provider supported by langchain-community.

Notice

Set a model by specifying its name in the config:

config.toml
ai_model = "gpt-4"

Supported Providers

Built-in (no extra packages needed):

  • OpenAI: gpt-4, gpt-3.5-turbo, etc.
  • Anthropic: claude-3-5-sonnet-20241022, claude-3-opus-20240229, etc.
  • Ollama: Custom models via ai_custom (see below)

Additional providers (require extra packages):

  • Google: Install langchain-google-genai for Gemini models
  • Others: Any LangChain-supported provider with the appropriate package

API Keys

Set the appropriate environment variable for your provider:

# OpenAI
export OPENAI_API_KEY="..."

# Anthropic
export ANTHROPIC_API_KEY="..."

# Google Gemini
export GOOGLE_API_KEY="..."

Custom/Self-Hosted Models (Ollama)

For self-hosted models like Ollama, define them in ai_custom:

[ai_custom."llama3.1:70b"]
server_type = "ollama"
url = "localhost:11434"
max_tokens = 128000

Then use it:

ai_model = "llama3.1:70b"

Installing Extra Provider Packages 

# Local installation
pip install langchain-google-genai

# With pipx
pipx inject esbmc-ai langchain-google-genai

# In containers (at build time)
hatch run podman-build "langchain-google-genai"
UsageBuilt-in Commands