v2.3.0

News

This release adds the new Condor MONIT toolset for CMS CompOps workflows, along with smaller configuration and deployment fixes for MCP servers, scraping, and packaged skills.

What's Changed

Condor Monitoring and CompOps Tooling

• Added new Condor MONIT OpenSearch tools to the CMS CompOps agent for searching and aggregating HTCondor job metrics.
• Added support for separate MONIT datasource configuration for Rucio and Condor endpoints.
• Included the new condor_raw_metric skill and improved MONIT aggregation fallback behavior for fields that do not expose .keyword buckets.

MCP Configuration and Runtime Support

• Fixed MCP server configuration loading to use dedicated mcp_servers config instead of reading from the general archi config.
• Persisted mcp_servers in seeded/static configuration and exposed it in generated deployment config.
• Improved MCP startup logging to show configured servers, loaded tools, and per-server fetch failures.

Scraping and Remote Selenium Support

• Added selenium_url configuration support so SSO scrapers can run against a remote Selenium driver.
• Improved scraper authentication and crawl logging with fuller exception details and more robust login element waits.

Deployment Packaging and Skills Mounting

• Updated deployment templating to copy configured skill markdown files into deployments.
• Mounted the packaged skills directory into runtime services so deployed chat applications can load those skills consistently.

Merged Pull Requests

• #485 fix-mcp-server-support
• #495 feature/remote_scraper
• #508 add-condor-os
• #520 fix_skills_mounting

Full Changelog: v2.2.0...v2.3.0

v2.2.0

News

This release focuses on reliability and observability improvements across chat, ingestion, and service operations. It has already been deployed to the production CompOps deployment to fix and smooth the experience for the users, with more updates and fixes in the pipeline, coming soon!

What's Changed

Service Operations and Alerting

• Added a new Service Status Board with alert management workflows.
• Introduced role-based permissions for alert manager access.
• Added dedicated service alert APIs and UI components for status visibility.

Timestamp and Data Consistency

• Migrated timestamp handling to timezone-aware UTC values across services.
• Updated SQL schema and API timestamp serialization to improve consistency.
• Added unit tests and follow-up compatibility fixes for timestamp behavior.

Providers and Model Visibility

• Added support for the CERN LiteLLM provider.
• Exposed the per-message model label in chat UI and persisted model-used values.
• Updated provider documentation and examples.

Chat and Conversation Reliability

• Improved context overflow handling in BaseReActAgent with retry/graceful fallback behavior.
• Fixed conversation user tracking by ensuring user_id is saved and queried consistently.
• Removed deprecated/debug auth and document-index code paths.

Scheduler, Upload, and UI Improvements

• Updated scheduler behavior to start even without preconfigured schedules.
• Improved upload and data-viewer frontend flows and related Playwright tests.
• Added and refreshed docs for service status board and alert configuration.

Merged Pull Requests

• #458 feat-service-status-board
• #477 fix/scheduler
• #482 fix-context-overflow
• #483 fix-to-save-userid
• #494 ui_fix
• #507 cernlitellm
• #511 fix_timestamp

Full Changelog: v2.1.0...v2.2.0

v2.1.0

News

This version is the current candidate for a first deployment to the extended CompOps team at CMS. Adds role-based access control alongside various tweaks and bug fixes on top of v2.0.0 -- see more below. Look out for additional features and fixes already under review to come soon in the next release.

What's Changed

Access Control and Permissions

• Introduced role-based access control (RBAC) and SSO-aligned role context handling.
• Improved permission validation and standardized checks using the Permission enum.
• Updated database viewer and related endpoints to use permission-based access control.
• Removed redundant/debug endpoints related to token and role inspection.

Configuration and Runtime Behavior

• Added configuration support to enable/disable providers.
• Added configurable chat timeout handling through app/config updates.
• Integrated role-context retrieval improvements in agent base tool paths.

UI and UX

• Updated login flow behavior to redirect to the landing page.
• Hid API Keys and Advanced settings sections in the chat UI -- require bug fixes + additional testing
• Updated associated UI tests to match the disabled/hidden settings behavior.

Cleanup and Refactors

• Refactored related app/tool code paths for consistency.

Merged Pull Requests

• #423 role-based-sso-auth
• #472 default-providers
• #473 chat-timeout-config
• #478 hassan11196-patch-2

Full Changelog: v2.0.0...v2.1.0

v2.0.0

Archi v2.0.0 Release Notes

Release date: February 25, 2026

v2.0.0 is a major release focused on backend consolidation, a redesigned chat experience, and operational flexibility for agent-driven workflows.

Highlights

Backend Consolidation to PostgreSQL + pgvector

• Consolidated core backend storage and retrieval on PostgreSQL, including vector search support with pgvector (PR #413).
• Removed legacy ChromaDB-oriented paths and related backend complexity as part of the consolidation.
• Improved ingestion and catalog reliability around Postgres behavior and constraints (PRs #420, #431, #466).

New Chatbot UI

• Shipped a significantly updated chatbot interface with stronger streaming UX, improved model selection flow, and better mobile/accessibility behavior (PR #413 stream).
• Improved agent activity visibility, including timeline/thinking-event oriented updates and token usage surfacing (PR #425).
• Patched tool rendering/availability issues in the chat tool UX (PR #445).

New Data Viewer and Uploader Experience

• Introduced consolidated data upload and data viewer pages for document workflows and data operations (PR #417).
• Improved document management behavior, loading robustness, extension handling, and large-file upload resilience (PRs #417, #437, #461).
• Added data exploration capabilities around metadata and chunk/document viewing in the new UI/backend flow.

Live Configurable Agents

• Added dynamic settings handling and agent configuration updates via markdown-driven settings, reducing restart-driven workflows (PR #426).
• Enabled dynamic schedule/config reload behavior to support live operational changes without full service restarts (included in v2.0.0 history).

Integrations and Tooling

• Added MCP integration and expanded tool ecosystem support (PR #415).
• Added MONIT OpenSearch tooling for Rucio event querying and skill-oriented workflows (PR #432).
• Improved provider handling and local provider configuration behavior (PRs #419, #436).

CI and Platform Improvements

• Migrated CI to GitHub runners and cleaned up CI/deployment behaviors for current workflows (PRs #438, #439).
• Added Windows-specific command parsing fix for better cross-platform CLI behavior (PR #437).

Breaking Changes

• Storage backend migration: this release formalizes a PostgreSQL-first backend (with pgvector) and removes legacy ChromaDB-era assumptions (PR #413).
• Configuration evolution: dynamic configuration and agent settings workflows have shifted, including markdown-based agent configuration support (PR #426).
• Provider/model path updates: provider/model unification and related config behavior may require existing deployments to align config keys and defaults (PR #419).

Migration Notes

1. Ensure PostgreSQL is provisioned and reachable by all relevant services, with pgvector enabled.
2. Validate ingestion/indexing configuration against the Postgres-backed data model before production rollout.
3. Review and migrate agent/settings configuration to the new dynamic/markdown-driven patterns where applicable.
4. Re-verify provider endpoints and local provider config (especially Ollama/local URLs) after upgrade.
5. Run end-to-end validation for:
   • chat streaming + tool invocation visibility
   • data uploader/viewer ingestion flow
   • agent activity traces/token metrics

Notable Fixes Included in v2.0.0

• Auto-trim conversation history based on provider context window to reduce overflow failures (PR #454).
• Improved unauthorized API response handling (PR #459).
• Better handling of recursion limit/error paths in agent workflows (PR #457).
• Data viewer loading/document-limit fixes (PR #461).
• Ingestion and scraping robustness improvements, including public-link and URL sanitization fixes (PRs #427, #431).

Scope Note

These notes are intentionally highlights-only (not an exhaustive commit/PR changelog) for the changes since v1.4.0.

v1.4.0

RELEASE NOTES:

• Data ingestion overhaul: persistence now catalogs resources in Postgres (CatalogService)
  instead of YAML indices; new data-manager Flask service with schedulable collectors,
  ingestion status API, and post‑ingestion vectorstore refresh (src/bin/
  service_data_manager.py, src/data_manager/*).
• Pipelines & agents: LangGraph-based ReAct agent layer added with DocumentMemory,
  configurable recursion limit, tool-call capture, and unified PipelineOutput (e.g.,
  base_react.py, agents/tools/*, utils/output_dataclass.py). Classic pipelines moved under
  pipelines/classic_pipelines/ with clearer module boundaries.
• Chat app upgrades: optional SSO/basic auth, per-config pipeline selection, conversation
  CRUD with metadata in Postgres, tool-call logging to agent_tool_calls, and UI
  refinements (src/interfaces/chat_app/app.py, src/utils/sql.py, base-init.sql).
• Retrieval & tools: built‑in agent tools for metadata search, grep-style content search,
  document fetch, and vectorstore search documented in docs/docs/user_guide.md; retriever
  config split into semantic/BM25/hybrid sections (base-config.yaml).
• Deployment/CLI: new archi restart command for single-service restarts; template drift
  warnings and Docker availability check; base compose adds data-manager service, mounts
  prompts/, updates Chroma healthcheck, and host‑mode/port handling improvements (src/cli/
  cli_main.py, base-compose.yaml, helpers.py).
• Benchmarking: query schema now supports explicit sources and RAGAS metrics; examples and
  docs refreshed (examples/benchmarking/queries.json, docs/docs/user_guide.md).
• Observability: Grafana dashboards/datasource templates updated; Postgres schema now
  includes agent_tool_calls plus conversation metadata timestamps (base-init.sql, grafana/
  *).
• CI/Dev tooling: new workflows for docs deploy, base image publish, PR preview; smoke
  test runner script; Docker image/tag management utilities; environment.yml removed,
  dependencies consolidated (.github/workflows/, scripts/dev/, environment.yml deleted).
• Rebranding & packaging: project/package renamed to archi, in preparation for full a2rchi --> archi migration leading up to v2.0.0.; version bumped to 1.2.4; new logo and GitHub home updated (pyproject.toml,
  README.md).

Breaking / Migration Notes

• Run pip install -e . (or reinstall) to pick up the renamed package/CLI and updated
  templates.
• Re-render configs/compose files with the new templates: chat app requires
  services.chat_app.pipeline; data manager now a separate service; retriever settings live
  under data_manager.retrievers.*; link lists are expected as relative files (e.g.,
  miscellanea.list).
• Apply DB init from the new base-init.sql to create conversation_metadata and
  agent_tool_calls tables before deploying updated services.
• Uploader service was dropped in favor of data-manager; compose/ports and secrets may
  need alignment.
• If using SSO/basic auth or Grafana, ensure env secrets exist and regenerate .env with
  the new options.

v1.2.0

A2rchi v1.2.0 Release Notes

Overview

This major release includes approximately 200+ commits since v1.1.0, bringing significant architectural improvements, new features, enhanced configurability, and better developer experience to A2RCHI.

Major Features

Multi-Configuration Support

• Multiple Prompt Configurations: Users can now run A2rchi with multiple configuration files simultaneously, allowing different prompt strategies for different use cases
• Dynamic Configuration Switching: New UI dropdown allows switching between different prompt configurations on the fly
• API Endpoint for Prompts: Added /prompts endpoint to retrieve available prompt configurations

Enhanced Data Sources & Scrapers

• Git Repository Scraper: New scraper for ingesting documentation directly from Git repositories, including support for MkDocs sites
  • Authentication support with username and personal access tokens
  • Automatic detection and handling of MkDocs repositories
  • Configurable via git_username and git_token secrets
• Sources Registry System: Implemented centralized sources registry for better management of data ingestion pipelines
• SSO Scraper Improvements: Enhanced SSO scraping with better recursion handling and URL tracking in vector database

Retrieval Enhancements

• Hybrid Search with BM25: Introduced hybrid retrieval combining semantic search with BM25 keyword matching for improved accuracy
  • Configurable hybrid retriever settings
  • Better handling of keyword-based queries
• Document Stemming: Optional stemming of documents before embedding creation for improved retrieval accuracy
  • Configurable stemming for both documents and queries
  • Particularly useful for technical documentation
• Embedding Options: Enhanced embedding configuration including:
  • Distance metric selection
  • Custom embedding instructions
  • Embedding model selection improvements

New LLM Integrations

• Ollama Support: Full integration with Ollama for local model inference
  • Configurable GPU allocation (num_gpu: -1 for all available GPUs)
  • Support for various Ollama models
• vLLM Improvements: Enhanced vLLM integration for faster inference
• Improved HuggingFace Support: Better handling of HuggingFace models for both inference and evaluation

Improvements

Architecture & Code Organization

• Chain/Pipeline Abstraction: Major refactoring of the chain system
  • Introduced BasePipeline for generalized LLM and prompt initialization
  • Support for multiple pipelines running simultaneously
  • Cleaner separation between chains, workflows, and wrappers
  • Created dedicated chains.py module
• Configuration Structure Overhaul: Completely restructured configuration file format
  • More intuitive hierarchy
  • Better validation and error handling
  • Support for pipelines (plural) in configuration

CLI Enhancements

• New CLI Implementation: Rebuilt CLI with improved functionality
  • --config flag to specify configuration files
  • -d/--dry-run mode for testing configurations without deployment
  • -f/--force flag for a2rchi create to automatically delete existing deployment first
  • --print-config option to display the loaded configuration
  • -p flag added to a2rchi delete for proper cleanup
• Better Secret Handling: Improved management of API keys and passwords across services

Container & Deployment

• Base Images: Created optimized base Docker images
  • pytorch-base and python-base images now available on DockerHub
  • Significantly faster deployment times
  • GPU and non-GPU variants for different use cases
• Slimmer Images: Optional lightweight images without GPU dependencies when running API-based models
• Requirements Reorganization: Split requirements into multiple files for better dependency management
• CUDA Version Update: Fixed CUDA mismatch issues, now running CUDA 12.4
• OpenShift/OKD Support: Added proper permissions and configurations for Kubernetes deployments
• Health Check Probes: Implemented health check endpoints for container orchestration

Network & API

• Host Mode Support: Fixed and improved host networking mode
  • Grafana works correctly in host mode
  • ChromaDB respects chromadb_external_port configuration in host mode
• Same-Origin API Calls: Optimized API calls from the frontend to avoid unnecessary host/port additions
• HTTPS Support: Frontend can now communicate with HTTPS APIs
• ChromaDB API Endpoints: Added REST API for ChromaDB operations
  • Document listing endpoint
  • Document search endpoint
  • Configurable enable/disable option
  • Comprehensive API documentation in user guide

Developer Experience

• GitHub Actions CI/CD:
  • Added smoke tests for PR validation
  • PR preview environment automation
  • Automated testing pipeline
• Logging Improvements:
  • Better structured logging across all containers

Interface Improvements

• Redmine Enhancements:
  • Bug fixes for Redmine mailer
  • Ticket client improvements
  • Better Postgres integration
• Grafana Monitoring:
  • Added retrieval scores to Grafana dashboards
  • Better visualization of context and history
  • Timeout and batch size configurations

Benchmarking & Evaluation

• Benchmarking Framework: New comprehensive benchmarking functionality
  • Support for multiple evaluation LLMs and providers
  • HuggingFace evaluation model support
  • Better default evaluation models
  • Configurable timeout, embedding model, verbosity, and batch_size
  • Queries configuration via queries.json
  • Plotting dependencies and visualization tools
  • Dedicated benchmarking documentation

Documentation

• Complete Documentation Overhaul: Comprehensive rewrite of user and developer guides
• New User Guide Sections:
  • ChromaDB API endpoints documentation
  • Hybrid search documentation
  • Stemming and Ollama interface documentation
  • Git scraper setup and usage
  • Benchmarking guide
  • Vector store configuration
• Developer Guide Updates: Enhanced developer documentation with architectural diagrams
• README Improvements: Updated README with new logo, clearer instructions, and current examples
• API Documentation: Complete API endpoint documentation
• Configuration Examples: Added example configurations for common use cases

Infrastructure Changes

• Directory Restructure: Major reorganization of src/data_manager/ and related directories
• Requirements Management: Split into multiple organized requirements files
• Automated Image Publishing: Script to push new Docker images to registry
• MkDocs Material: Integration with MkDocs Material for enhanced documentation
• Firefox GPU Support: Fixed Firefox compatibility for GPU-accelerated instances
• .gitignore Updates: Proper handling of .env files and .github workflows

Configuration Changes

New Configuration Options

• chromadb_external_port: Configure external port for ChromaDB in host mode
• enable_chromadb_api: Toggle ChromaDB API endpoints
• hybrid_search: Enable/disable hybrid retrieval
• stemming: Configure document and query stemming
• distance_metric: Choose embedding distance metric
• embedding_instructions: Custom instructions for embedding models
• ollama_num_gpu: GPU allocation for Ollama models
• timeout, batch_size, verbosity: Benchmarking configurations
• pipelines: Support for multiple pipeline configurations (replaces single chain)

Breaking Changes

• Configuration structure has been significantly refactored
• chain configuration section renamed to pipelines (plural)
• Some configuration keys have been reorganized into new hierarchies
• Old configuration files will need to be migrated to new format

Testing & Quality

• Added smoke tests for core functionality
• Implemented GitHub Actions for automated testing
• PR preview environments for testing changes
• Improved error handling throughout codebase
• Better validation of configuration files

Resources

• [Documentation](https://mit-submit.github.io/A2rchi/)
• [User Guide](https://mit-submit.github.io/A2rchi/user_guide/)
• [GitHub Repository](https://github.com/mit-submit/A2rchi)
• [Issue Tracker](https://github.com/mit-submit/A2rchi/issues)

---

Full Changelog: v1.1.0...v1.2.0

v1.1.0

v1.1.0 Release Notes

Performance:

• vLLM implemented for orders of magnitude faster inference when running open source models on GPU. Control how many GPUs to run distributed on, or how much memory to allocate to vLLM, and more, all via the configuration file
• Better and more robust handling of chat history so A2rchi "remembers" conversations more accurately
• Host mode option for container networking

Interfaces and Integrations:

• Scraper to get behind CERN SSO (+ base SSO class to build other SSO scrapers on)
• Jira interface added to store tickets from a specified URL and project into vector database
• Redmine service integrated with Postgres and Grafana monitoring
• Vanilla Mattermost interface has been implemented
• Grafana monitoring upgraded, including now formatted history and context, amongst other changes

General:

• Various bug fixes and stability improvements
• User guide updated

Developer:

• Improved and uniformized logging in containers
• Templated prompts and LLM outputs now nicely organized in chain_input_output_log, stored in container volume, for studies/further debugging

v1.0.0

Removed CI/CD pipeline, A2rchi runs via a CLI now. See more of what A2rchi provides at https://mit-submit.github.io/A2rchi/

v0.1-gh-cicd: Merge pull request #169 from mit-submit/cpausmit-patch-1

old version of a2rchi framework that used github actions for CI/CD pipeline.