Skip to main content

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.

IssueLoop configuration is passed to issueloop.use() once at startup, before any scanning or ticket operations. Every option has a sensible default, so zero configuration is needed for a local trial — a fresh install works out of the box with a local Ollama instance.

​
The use() function

Call use() at the top of your script (or in your application’s startup routine) to configure the library globally. Calling it a second time replaces the previous configuration entirely. 
issueloop.use(
    database="local",
    database_path=None,
    retention_days=30,
    llm=None,
    notify=None,
)
Each keyword argument maps to a top-level configuration concern. Arguments you omit fall back to their defaults. The function returns the newly constructed Config object, though you rarely need that return value — get_config() retrieves it at any time.

​
All configuration options

The table below covers every accepted field, including the nested keys inside the llm and notify dicts.
FieldValuesDefaultDescription
database"local", "supabase""local"Storage backend. "local" uses a SQLite file; "supabase" uses a hosted Postgres-backed project.
database_pathany filesystem pathNone (resolves to data/issueloop.db at runtime)Path to the local SQLite database. When None, the backend applies data/issueloop.db relative to the IssueLoop package root. Ignored when database="supabase".
retention_daysinteger30How many days completed tickets are kept before cleanup removes them.
llm.providerslist of provider dicts, in priority ordersingle Ollama defaultPriority-ordered list of LLM backends. IssueLoop falls back to the next entry automatically on failure.
llm.provider"ollama", "anthropic", "openai""ollama"LLM provider for single-provider configurations.
llm.apiKeystringNone (required for anthropic/openai)API key for the selected provider.
llm.tokenSizeinteger1024Maximum token budget passed to the LLM per call.
notify.webhookURL stringNoneEndpoint that receives a POST request if IssueLoop itself encounters an internal error.

​
LLMConfig fields

Internally, each entry in llm.providers (or the single llm dict) is normalised into an LLMConfig dataclass. The fields and their defaults are:
​
provider
str
default:"ollama"
The LLM backend to use. Accepted values are "ollama", "anthropic", and "openai".
​
model
str
default:"qwen2.5-coder:7b"
The model name as the provider expects it. For Ollama this is the tag used with ollama pull; for Anthropic and OpenAI it is the model identifier from their respective APIs.
​
api_key
str | None
default:"None"
API key for cloud providers. Not required for a local Ollama instance. Accepts both apiKey (camelCase) and api_key (snake_case) in the raw dict passed to use().
​
base_url
str
default:"http://localhost:11434"
Base URL for the provider’s API. The default points to a locally running Ollama server. Override this for a remote Ollama instance or a custom OpenAI-compatible endpoint.
​
token_size
int
default:"1024"
Maximum tokens per LLM call. Accepts both tokenSize (camelCase) and token_size (snake_case) in the raw dict.

​
NotifyConfig

The notify argument accepts a plain dict. Both fields are optional.
​
email
str | None
default:"None"
Email address to notify on internal IssueLoop errors.
​
webhook
str | None
default:"None"
HTTP/HTTPS URL that receives a POST payload when IssueLoop encounters a crash or unhandled exception.

​
camelCase and snake_case

IssueLoop accepts both naming conventions in the dicts you pass to use(). The internal _build_llm_config function normalises them:
  • apiKey and api_key are both read; apiKey is tried first.
  • tokenSize and token_size are both read; tokenSize is tried first.
  • baseUrl and base_url are both read; baseUrl is tried first.
This means the following two calls are equivalent: 
# camelCase style
issueloop.use(llm={"provider": "openai", "apiKey": "sk-...", "tokenSize": 2048})

# snake_case style
issueloop.use(llm={"provider": "openai", "api_key": "sk-...", "token_size": 2048})
Mix and match freely — whichever key is present wins.

​
Config resolution order

get_config() always returns the current global config object. Calling use() replaces it atomically for all subsequent operations. 
import issueloop

# Before any use() call — defaults are active
cfg = issueloop.get_config()
print(cfg.database)          # "local"
print(cfg.retention_days)    # 30

# Reconfigure
issueloop.use(database="supabase", retention_days=7)

# Now the new config is in effect everywhere
cfg = issueloop.get_config()
print(cfg.database)          # "supabase"
print(cfg.retention_days)    # 7
There is no file-based configuration for these options — they are code-only. The provider_config.yaml file is a separate convenience layer for the LLM backend only; see Provider Config for details.
use() is entirely optional. The defaults work out of the box with a local Ollama install — no API key, no database account, and no config file required to run your first scan.

Build docs developers (and LLMs) love

Get started for freeTalk to us