RepoGhost: AI Sidecar for Engineering Output
Quick Summary
- AI-powered engineering output generation
- Integrates meeting context and repo data
- Local or hosted deployment options
- Automates ticket, standup, PR tasks
- Focuses on data privacy and control
RepoGhost is an innovative graph-native AI sidecar designed to enhance engineering productivity by converting fragmented meeting contexts and complex repository structures into clear, actionable outputs. This intelligent digital assistant works alongside your existing tools, streamlining development workflows. It helps teams move from raw signals to concrete engineering artifacts, making the often-tedious process of documentation and task creation more efficient. Whether you prefer a local setup or a hosted application, RepoGhost offers flexibility in deployment. It aims to bridge the gap between team discussions, code changes, and project management tasks, allowing engineers to focus more on development and less on administrative overhead.
- 📋 Setting Up Your Workspace for RepoGhost Integration
- 📍 Initiating Your RepoGhost AI Sidecar Session
- ⭐ Transforming Code Context into Actionable Engineering Outputs
- ⚠️ Addressing Workflow Interruptions with RepoGhost
- 🔒 Maintaining Confidentiality and Control Over Your Repository Data
- 🎧 Official Help Desk Contacts
- ❓ Your Top Questions About RepoGhost Answered
Setting Up Your Workspace for RepoGhost Integration
- Docker installed on your system for local deployment (for x86/amd64 or ARM64 architectures).
- A supported operating system: macOS, Windows, or Linux (native builds are coming soon, Docker is current method).
- For building from source: Jac runtime installed.
- Sufficient system resources to run Docker and the RepoGhost application efficiently alongside your Integrated Development Environment (IDE).
Initiating Your RepoGhost AI Sidecar Session
- To access the hosted application: Navigate to the RepoGhost website and look for the option to Try the hosted app instantly.
- For local deployment via Docker: First, ensure Docker is installed and running on your machine.
- Open your terminal or command prompt.
- For x86/amd64 systems, run the command: `docker pull ghcr.io/Async-Avengers/repoghost:latest`
- Then, execute: `docker run -p 8000:8000 -p 8001:8001 repoghost -v /home:/home` (adjust `/home` to your desired volume mount if necessary).
- For ARM64 systems, run: `docker pull ghcr.io/async-avengers/repoghost:latest-arm`
- Then, execute: `docker run -p 8000:8000 -p 8001:8001 repoghost -v /home:/home`.
- Once running, open your web browser and navigate to http://localhost:8001 to begin using RepoGhost locally.
Transforming Code Context into Actionable Engineering Outputs
- Three-Stage AI Pipeline: RepoGhost operates through a sophisticated three-stage process to deliver actionable engineering insights.
- Intelligent Ingestion: Utilizes six specialized agents including RepoScanner, TranscriptLoader, DiffAnalyzer, DocCrawler, IssueMapper, and HistoryWalker to gather data from various sources like repositories, meetings, diffs, and documentation.
- Contextual Reasoning: Employs four reasoning agents: ContextLinker, DepResolver, RiskScorer, and IntentMapper, which cross-reference contextual nodes, identify dependencies, and annotate the graph with engineering intent.
- Automated Output Generation: Features five output agents – TicketWriter, StandupGen, HandoffDraft, PRScopeGen, and ReviewAnnotator – to synthesize complex data into concrete engineering artifacts.
- Diverse Engineering Artifacts: Generates practical outputs such as GitHub Tickets, Standup Drafts, Handoff Docs, and PR Review Scopes to streamline project management.
- Local Execution Option: Can be run locally using Docker, providing an offline capability and enhanced data control without requiring a connection to cloud services.
Addressing Workflow Interruptions with RepoGhost
⚠️ RepoGhost fails to start after Docker command
Cause: Often, this is due to Docker not running or incorrect Docker volume mapping.
Solution: Ensure your Docker Desktop application is actively running. Verify that the volume mount path (`-v /home:/home` in the command) is correctly configured for your operating system and desired local directory. Check the Docker logs for specific error messages.
⚠️ Unable to access RepoGhost in browser after starting Docker
Cause: The local server might not be accessible on the specified port, or another application is using port 8001.
Solution: Confirm that you are navigating to http://localhost:8001. If the issue persists, check if another process is utilizing port 8001. You can often change the port mapping in the `docker run` command if necessary.
⚠️ Issues when building RepoGhost from source
Cause: Missing the Jac runtime or incorrect repository cloning.
Solution: Ensure Jac runtime is installed as a prerequisite. Confirm that you have correctly cloned the repository using `git clone https://github.com/Async-Avengers/repoGhost.git` and then run `jac install` followed by `jac build; ./run.sh`.
Maintaining Confidentiality and Control Over Your Repository Data
- 🔒 Local Deployment for Enhanced Privacy: For sensitive projects, leverage the local Docker deployment option to keep your code context and generated engineering outputs entirely within your own environment, as it states "no cloud required" for this setup.
- 🔒 Secure Your Local Machine: Always ensure your local development machine, where RepoGhost is running, is secured with up-to-date antivirus, a strong firewall, and regular security patches.
- 🔒 Control Data Access: Be mindful of the repositories and meeting transcripts you provide to RepoGhost, especially when using the hosted application. Understand what data is being ingested.
- 🔒 Review Generated Outputs: While RepoGhost aims for accuracy, always review the AI-generated engineering artifacts (tickets, PR scopes, etc.) before integrating them into your official project workflows.
🎧 Official Support & Help Center
No specific contact support phone number or email is directly listed on the main page of repo-ghost.com. The product is built by Async Avengers and powered by Jac & byllm. For support, users might need to explore deeper links on the website or look for community forums/GitHub repositories associated with the project.
Your Top Questions About RepoGhost Answered
Q: What kind of data does RepoGhost use?
RepoGhost ingests data from various sources, including your code repository, meeting transcripts, code differences (diffs), documentation, issue trackers, and historical data, to build a comprehensive context graph.
Q: Can I run RepoGhost without an internet connection?
Yes, if you deploy RepoGhost locally using Docker, it can run without a constant internet connection, as it's designed as a desktop sidecar app that operates locally and states "no cloud required" for this setup.
Q: What types of engineering tasks can RepoGhost automate?
RepoGhost can automate the creation of various engineering artifacts such as GitHub Tickets, Standup Drafts, Handoff Documents, and Pull Request Review Scopes.
Q: Who developed RepoGhost?
RepoGhost was built by a team known as Async Avengers and is powered by Jac and byllm technologies.
Wrap Up
RepoGhost offers a compelling solution for engineering teams looking to enhance their productivity by intelligently transforming disparate information into structured, actionable engineering outputs. Its ability to integrate with existing repositories and communication channels, combined with flexible deployment options, positions it as a valuable AI sidecar for modern software development. By automating routine documentation and task creation, RepoGhost empowers engineers to streamline their workflows and dedicate more time to core development efforts.