Changelog

Release history

We ship updates roughly every two to four weeks. Below is the full release history since our first beta in mid-2024.

  • v0.9

    CI/CD integration and full API access

    • New REST API to trigger eval runs from CI pipelines (Pro plan)
    • New GitHub Actions workflow example in the docs
    • Improvement Deployment API now returns structured errors for failed eval gates
    • Fix Fixed a race condition when two team members trigger one eval run at once
  • v0.8

    Regression checks and prompt diff viewer

    • New Automatic regression checks compare new eval results with a pinned baseline
    • New Compare any two prompt versions side by side
    • Improvement Eval history now includes per-case scores, not only aggregate pass/fail
  • v0.7

    Team workspaces and shared rubrics

    • New Multi-member workspaces with role-based access (Admin, Editor, Viewer)
    • New Shared evaluation rubrics for every team member
    • Improvement Projects now show the last editor and eval date
    • Fix Fixed cached prompt renders not updating after variable schema changes
  • v0.6

    Staged deployment rollouts

    • New Rollout controls: deploy to 10%, 25%, 50%, or 100% of traffic
    • New Deployment timeline with rollback to any earlier version
    • Improvement Confirmation now shows the eval gate summary before deployment
  • v0.5

    Custom eval rubric builder

    • New Build custom rubrics with natural-language criteria and numeric scoring thresholds
    • New Rubric templates for classification accuracy, tone adherence, and factual grounding
    • Fix Fixed UI stalls caused by eval cases with long model outputs
  • v0.4

    Python SDK and TypeScript client

    • New Python SDK: pip install monogramml for programmatic prompt and eval access
    • New TypeScript client for Node.js and browser environments
    • Improvement SDK offers pagination helpers for large eval result sets
  • v0.3

    Eval engine and run history

    • New Structured evals: define cases with expected outputs and scoring functions
    • New Run history: timestamps, per-case pass/fail, and aggregate score
    • Improvement Prompt workspace now shows the eval suite linked to each prompt
  • v0.2

    Prompt versioning and variable schemas

    • New Named prompt versions with commit-style labels and full history
    • New Typed variable injection schemas: string, number, list, JSON object
    • Fix Fixed multi-turn conversation prompts rendering in the preview panel
  • v0.1

    Private beta launch

    • New Prompt workspace with basic version tracking
    • New Project-level organization for multiple prompt sets
    • New Manual deployment for pushing prompt versions to a staging endpoint