These functions handle initial configuration, system health checks, and diagnostics.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/onenot8/issueLoop/llms.txt
Use this file to discover all available pages before exploring further.
use is the only function you must call before the rest of the API — every other function reads the global config object that use populates. All other functions in this group are read-only and safe to call at any time.
use
The primary configuration function. Call it once at startup, before any otherissueloop function. Calling it a second time replaces the entire global config.
Storage backend to use.
"local" stores tickets in a local SQLite file. "supabase" uses a Supabase-hosted database.Path to the SQLite database file. Only used when
database="local". Defaults to data/issueloop.db relative to the IssueLoop working directory.Default age threshold for
cleanup(). Tickets older than this many days are eligible for deletion by cleanup() when no explicit older_than_days is passed.LLM provider configuration. Pass a single provider dict for one provider, or wrap multiple in
When omitted, IssueLoop falls back to Ollama with
{"providers": [...]} for automatic fallback. Each provider dict supports:| Key | Description | Default |
|---|---|---|
provider | "ollama", "anthropic", or "openai" | "ollama" |
model | Model name | "qwen2.5-coder:7b" |
api_key / apiKey | API key (required for Anthropic and OpenAI) | none |
base_url / baseUrl | Base URL for Ollama or a compatible endpoint | "http://localhost:11434" |
token_size / tokenSize | max_tokens for Anthropic/OpenAI calls | 1024 |
qwen2.5-coder:7b at http://localhost:11434.Notification configuration. Errors caught internally by IssueLoop (LLM failures, escalated tickets) are sent here.
| Key | Description |
|---|---|
webhook | URL to POST error payloads to |
email | Email address (reserved for future use) |
Config object (a dataclass with database, database_path, retention_days, llm, llm_providers, and notify fields).
Calling
use() a second time replaces the global config in its entirety. If you need to update a single setting, call get_config() first, extract the values you want to keep, then call use() with the full set.get_config
Returns the current globalConfig object as set by the most recent use() call. Useful for inspecting active settings or for passing config values to lower-level modules.
Config dataclass instance. Before use() is ever called, returns a default Config with database="local", retention_days=30, and Ollama as the LLM provider.
list_repos
Returns the list of repository names defined indata/test_manifest.json. Returns an empty list if the manifest file does not exist.
list[str] of repository name strings in the order they appear in the manifest.
health_check
Runs a basic liveness check on the configured backend and returns a summary dict. Does not run tests or query tickets — it only verifies that the database connection can be established.The configured backend type.
The name of the primary LLM provider.
True if all checks passed. False if any check raised an exception.List of human-readable error messages for any failed checks. Empty when
ok is True.get_crash_log
Returns IssueLoop’s own internal error log — entries written whenever the library catches an unexpected exception (LLM provider failures, escalated tickets, notification delivery errors).Maximum number of entries to return. Returns the most recent
limit entries from data/logs/crashes.jsonl.ISO 8601 UTC timestamp of when the error was caught.
A short label identifying which internal operation failed (e.g.
"llm.chat (provider 1/2: anthropic)", "ticket.escalate").String representation of the exception.
Full Python traceback as a string.
data/logs/crashes.jsonl does not yet exist.
get_notification_config
Returns the notification settings from the current config without exposing the fullConfig object.
Configured webhook URL, or
None if not set.Configured email address, or
None if not set.