Configuration Reference

Quick reference for all config.toml options. For detailed explanations, see the linked Systems pages.

Global Settings

workspace = "~/.ash/workspace"

Option	Type	Default	Description
workspace	path	"~/.ash/workspace"	Directory for SOUL.md and skills

API Keys

[anthropic]
api_key = "sk-ant-..."

[openai]
api_key = "sk-..."

Section	Option	Env Variable	Description
[anthropic]	api_key	ANTHROPIC_API_KEY	Anthropic API key
[openai]	api_key	OPENAI_API_KEY	OpenAI API key

Models

See LLM Providers for details.

[models.default]
provider = "anthropic"
model = "claude-haiku-4-5"
temperature = 0.7
max_tokens = 4096

[models.sonnet]
provider = "anthropic"
model = "claude-sonnet-4-5"
max_tokens = 8192

Option	Type	Default	Description
provider	string	required	"anthropic" or "openai"
model	string	required	Model identifier
temperature	float	null	Sampling temperature (0.0-1.0)
max_tokens	int	4096	Maximum response tokens
thinking	string	null	Extended thinking: "off", "minimal", "low", "medium", "high"

Sandbox

See Sandbox for details.

[sandbox]
image = "ash-sandbox:latest"
timeout = 60
memory_limit = "512m"
cpu_limit = 1.0
runtime = "runc"
network_mode = "bridge"
dns_servers = []
http_proxy = ""
workspace_access = "rw"
sessions_access = "none"

Option	Type	Default	Description
image	string	"ash-sandbox:latest"	Docker image name
timeout	int	60	Command timeout in seconds
memory_limit	string	"512m"	Container memory limit
cpu_limit	float	1.0	CPU cores allowed
runtime	string	"runc"	Container runtime ("runc" or "runsc")
network_mode	string	"bridge"	Network mode ("none" or "bridge")
dns_servers	list	[]	Custom DNS servers
http_proxy	string	""	HTTP proxy URL
workspace_access	string	"rw"	Workspace mount ("none", "ro", "rw")
sessions_access	string	"none"	Sessions mount ("none", "ro")

Memory

See Memory for details.

[memory]
database_path = "~/.ash/memory.db"
max_context_messages = 20
context_token_budget = 100000
recency_window = 10
system_prompt_buffer = 8000
auto_gc = true
max_entries = null
compaction_enabled = true
compaction_reserve_tokens = 16384
compaction_keep_recent_tokens = 20000
compaction_summary_max_tokens = 2000
extraction_enabled = true
extraction_model = null
extraction_min_message_length = 20
extraction_debounce_seconds = 30
extraction_confidence_threshold = 0.7

Option	Type	Default	Description
database_path	path	"~/.ash/memory.db"	SQLite database path
max_context_messages	int	20	Maximum messages in context
context_token_budget	int	100000	Target context window size
recency_window	int	10	Always keep last N messages
system_prompt_buffer	int	8000	Reserved tokens for system prompt
auto_gc	bool	true	Run garbage collection on startup
max_entries	int	null	Cap on active memories
compaction_enabled	bool	true	Enable context compaction
compaction_reserve_tokens	int	16384	Buffer before triggering
compaction_keep_recent_tokens	int	20000	Always keep recent context
compaction_summary_max_tokens	int	2000	Max tokens for summary
extraction_enabled	bool	true	Enable auto memory extraction
extraction_model	string	null	Model for extraction
extraction_min_message_length	int	20	Skip short messages
extraction_debounce_seconds	int	30	Min seconds between extractions
extraction_confidence_threshold	float	0.7	Minimum confidence for storing

Embeddings

See Memory for details.

[embeddings]
provider = "openai"
model = "text-embedding-3-small"

Option	Type	Default	Description
provider	string	"openai"	Embedding provider
model	string	"text-embedding-3-small"	Embedding model

Skills

See Skills for details.

[skills]
auto_sync = true
update_interval = 24

[[skills.sources]]
repo = "anthropic/skills"

[[skills.sources]]
path = "~/my-local-skills"

[skills.research]
enabled = true
PERPLEXITY_API_KEY = "pplx-..."

Global Settings

Option	Type	Default	Description
auto_sync	bool	false	Sync sources on startup
update_interval	int	24	Hours between auto-updates

Source Fields

Field	Type	Description
repo	string	GitHub repo in owner/repo format
path	string	Local filesystem path
ref	string	Git ref (branch, tag, commit)

Per-Skill Fields

Field	Type	Description
enabled	bool	Enable/disable the skill
model	string	Model alias override
*	string	Environment variables (auto-uppercased)

Agents

See Agents for details.

[agents.research]
model = "sonnet"
max_iterations = 50

[agents.skill-writer]
model = "sonnet"
max_iterations = 20

Option	Type	Default	Description
model	string	null	Model alias override
max_iterations	int	varies	Maximum iterations

Telegram

See Providers for details.

[telegram]
bot_token = "123456789:ABC..."
allowed_users = ["@yourusername", "123456789"]
allowed_groups = ["-100123456789"]
group_mode = "mention"
webhook_url = "https://your-domain.com/webhook"

Option	Type	Default	Description
bot_token	string	required	Bot token from BotFather
allowed_users	list	[]	Authorized usernames or IDs
allowed_groups	list	[]	Authorized group chat IDs
group_mode	string	"mention"	"mention" or "always"
webhook_url	string	null	Webhook URL (polling if not set)

Server

See Providers for details.

[server]
host = "127.0.0.1"
port = 8080
webhook_path = "/webhook"

Option	Type	Default	Description
host	string	"127.0.0.1"	Bind address
port	int	8080	Port number
webhook_path	string	"/webhook"	Telegram webhook path

Sessions

See Providers for details.

[sessions]
mode = "persistent"
max_concurrent = 2

Option	Type	Default	Description
mode	string	"persistent"	"persistent" or "fresh"
max_concurrent	int	2	Maximum parallel sessions

Brave Search

[brave_search]
api_key = "BSA..."

Option	Type	Default	Description
api_key	string	required	Brave Search API key

Getting an API Key

1. Sign up at Brave Search API

2. Create an API key in the dashboard

3. Configure Ash:

   [brave_search]
   api_key = "BSA..."

   Or use environment variable: BRAVE_SEARCH_API_KEY

Sentry

[sentry]
dsn = "https://abc123@o123.ingest.sentry.io/456"
environment = "production"
release = "1.0.0"
traces_sample_rate = 0.1
profiles_sample_rate = 0.0
send_default_pii = false
debug = false

Option	Type	Default	Description
dsn	string	null	Sentry DSN
environment	string	null	Environment name
release	string	null	Release version
traces_sample_rate	float	0.1	Transaction sampling (0.0-1.0)
profiles_sample_rate	float	0.0	Profiling sampling (0.0-1.0)
send_default_pii	bool	false	Include PII in reports
debug	bool	false	Enable debug logging

Caution

Set send_default_pii = false to avoid sending personal information to Sentry.

Workspace Files

The workspace directory contains:

~/.ash/workspace/
├── SOUL.md      # Assistant personality
├── USER.md      # User profile (optional)
└── skills/      # Custom skills

SOUL.md

Defines your assistant’s personality:

# Ash

You are a personal assistant named Ash.

## Personality
- Helpful and direct
- Technical but accessible
- Concise responses

## Guidelines
- Always verify before executing destructive commands

USER.md (optional)

Describes the user for personalized responses:

# User Profile

## About
- Software engineer
- Works on Python and TypeScript projects

## Preferences
- Concise code examples
- Unix command line tools