Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/clyrisai/gitresolve/llms.txt

Use this file to discover all available pages before exploring further.

GitResolve supports three output modes that can be used individually or in combination. The default terminal display is great for interactive use, --json makes it easy to pipe results into other tools, and --output-dir writes clean per-candidate files suitable for downstream automation.

​
Terminal Display (Default)

When neither --json nor --output-dir is provided, GitResolve prints an ANSI-colored summary table to your terminal. Each processed candidate gets one row with their resolved username, confidence level, number of owned repos, contributions, and any warnings. A summary line is printed at the end showing total resolved and unresolved counts. 
gitresolve --all

🔍 ClyrisAI GitResolve
────────────────────────────────────────

📂 Portfolio Links (from ./data/portfolio_links.csv)
   Found 3 URL(s)
   ✅ janedoe          [high]     3 repos  1 contribution
   ✅ jsmith           [medium]   1 repo   0 contributions
   ⚠️  alice-lee        [low]      0 repos  0 contributions

📂 Resumes (from ./data/resumes/)
   Found 2 PDF(s)
   ✅ janedoe          (merged with portfolio source)
   ✅ bob-chen         [high]     5 repos  2 contributions

────────────────────────────────────────
Resolved: 3   Unresolved: 1
The terminal display is suppressed when --json is active, keeping stdout clean for piping.

​
--json Mode

The --json flag suppresses all terminal UI output and prints the full results as a JSON array to stdout. Each element is an AggregatedResult object. This is the recommended mode when feeding GitResolve output into another tool or script. 
gitresolve --all --json
gitresolve https://github.com/janedoe --json

[
  {
    "candidateUsername": "janedoe",
    "sources": ["https://janedoe.dev", "./data/resumes/janedoe.pdf"],
    "sourceTypes": ["portfolio", "resume_file"],
    "ownerProfile": {
      "url": "https://github.com/janedoe",
      "provider": "github",
      "type": "profile",
      "username": "janedoe"
    },
    "confidence": "high",
    "ownedRepos": [
      {
        "url": "https://github.com/janedoe/my-project",
        "provider": "github",
        "type": "repo",
        "username": "janedoe",
        "repo": "my-project"
      }
    ],
    "contributions": [],
    "externalRepos": [],
    "allLinks": [],
    "warnings": []
  }
]
Combine --json with jq for quick field extraction:
# Extract all resolved usernames
gitresolve --all --json | jq '[.[] | .candidateUsername] | map(select(. != null))'

# Get all owned repo URLs for a specific user
gitresolve https://github.com/janedoe --json | jq '.[0].ownedRepos[].url'

# Filter to high-confidence results only
gitresolve --all --json | jq '[.[] | select(.confidence == "high")]'

​
--output-dir Mode

The --output-dir <dir> flag writes one JSON file per candidate to a directory on disk. GitResolve creates two subdirectories automatically:
  • <dir>/resolved/ — one <username>.json file per successfully resolved candidate
  • <dir>/unresolved/ — one <sanitized_source>_N.json file per candidate that could not be resolved
gitresolve --all --output-dir ./results

​
Output Directory Structure

results/
├── resolved/
│   ├── janedoe.json
│   ├── jsmith.json
│   └── bob-chen.json
└── unresolved/
    └── alice_lee_pdf_1.json
Each file contains the full AggregatedResult JSON for that candidate. Unresolved filenames are derived from the first source’s basename with non-alphanumeric characters replaced by underscores, followed by an incrementing counter.
The resolved/ and unresolved/ subdirectories are created automatically. You do not need to create them in advance.

​
Combining --json and --output-dir

Both flags can be used together. When combined, GitResolve writes per-candidate JSON files to disk and prints the full JSON array to stdout simultaneously. This is useful for saving a permanent archive while also piping the results into a downstream script in one step. 
gitresolve --all --json --output-dir ./results | jq '[.[] | .candidateUsername]'
The two outputs are independent — the files on disk are not affected by what you pipe stdout into.

​
AggregatedResult Schema

Every output mode produces AggregatedResult objects. The table below describes every field.
FieldTypeDescription
candidateUsernamestring | nullResolved Git username, or null if no profile could be identified
sourcesstring[]All input sources (URLs and file paths) that contributed to this result
sourceTypesInputType[]Classification of each source: portfolio, git_profile, repo_url, resume_file, resume_url, linkedin, or unknown
ownerProfileExtractedGitLink | nullThe best-guess owner profile link, or null if unresolved
confidence"high" | "medium" | "low" | "none"How confident GitResolve is that the correct owner was identified
ownedReposExtractedGitLink[]Repository links where the owner username matches candidateUsername
contributionsExtractedGitLink[]Explicit pull request and issue links attributed to the candidate
externalReposExtractedGitLink[]Repository links owned by other users (external references found on the page)
allLinksExtractedGitLink[]All Git links extracted from all sources, unfiltered
warningsstring[]Non-fatal messages such as unresolvable input types or scraping errors
Each link object inside ownedRepos, contributions, externalRepos, and allLinks has the following shape:
FieldTypeDescription
urlstringThe normalized URL of the link
provider"github" | "gitlab" | "bitbucket"The Git hosting provider
type"profile" | "repo" | "gist" | "pull_request" | "issue" | "other"The kind of Git link
usernamestringThe owner or author username extracted from the URL path
repostring (optional)Repository name, present when type is repo, pull_request, or issue
numberstring (optional)PR or issue number, present when type is pull_request or issue

​
Confidence Levels

ValueMeaning
highA Git profile URL was the direct input, or a single username dominated all discovered links
mediumA username was inferred from multiple links but with some ambiguity
lowA username was found but evidence was sparse or conflicting
noneNo Git profile could be identified from the input
A confidence of "none" does not always indicate an error. LinkedIn URLs, unknown input types, and pages with no Git links all produce "none" by design. Check the warnings array for a human-readable explanation.

Build docs developers (and LLMs) love

Get started for freeTalk to us