Thanks to visit codestin.com
Credit goes to Github.com

Skip to content

Releases: TonyStef/Grov

Git for Memories

22 Jan 02:43

Choose a tag to compare

0.6.15

change ytb link location

Agent expansion

09 Jan 16:33

Choose a tag to compare

fix: sync in postinstall

old auto sync in postinstall didn't work because of the interactive nature (sync y/n) -> postinstall is non interactive => solution: auto sync without any input from the user

Intelligent Memory Updates

22 Dec 14:49

Choose a tag to compare

Memories now evolve instead of duplicating. When you work on the same problem across sessions, Grov updates the existing memory rather than creating redundant entries.
How it works:

  1. Session ends → reasoning + decisions extracted
  2. Hybrid search (semantic + lexical) finds related memories
  3. AI decides whether to update or create new:
    • Update: If files were modified OR decisions contradict previous ones
    • Skip: Pure Q&A with no code changes
    • New: No matching memory found

What gets merged:

  • New reasoning appended to trace
  • Conflicting decisions supersede old ones
  • Full history preserved in evolution_steps

Dashboard Changes

  • Evolution Timeline: See how a memory changed over time
  • Previous Reasoning: View superseded reasoning entries
  • "Edited" badge: Memories show when they were last modified vs created
  • Decisions list: Shows which decisions are current vs superseded

Technical Details

  • Memories split into ~10-15 semantic chunks for better matching
  • Chunk types: summary, query, goal, reasoning entries, decisions
  • Search finds best chunk → returns parent memory

0.5.9 Fix grov --version command

19 Dec 14:44

Choose a tag to compare

fix grov --version command

Windows Support, Grov Uninstall, Better Documentation

19 Dec 14:38

Choose a tag to compare

feat: add grov uninstall command + Windows support for API key setup

- Add grov uninstall command with detailed confirmation
- Add Windows detection in grov init (shows setx instead of export)
- Add API Key Setup section to docs with OS-specific instructions
- Add permanent API key warning to README Quick Start
- Bump version to 0.5.8

v0.5.7 - Better onboarding & setup guidance

19 Dec 14:04

Choose a tag to compare

What's New
Improved Setup Guidance
Users were signing up but memories weren't syncing because API keys weren't persisting across terminal sessions. This release adds clear warnings and fix instructions at every step.

  • grov init - Now shows a prominent warning if ANTHROPIC_API_KEY is not set, with instructions to add it permanently to your shell profile
  • grov login - Warns after login if API key is missing
  • grov doctor - Improved output with copy-paste fix commands
    Documentation
  • Added Troubleshooting section to README
  • Added grov doctor command to docs and README
  • Updated pricing page with new tiers

Fix
If your memories aren't syncing, run:
grov doctor
If API key shows as not set, add it permanently:
echo 'export ANTHROPIC_API_KEY=sk-ant-...' >> ~/.zshrc
source ~/.zshrc

Add Grov Doctor Command

19 Dec 12:51

Choose a tag to compare

add: grov doctor command & bump version to 0.5.6

-Change pricing

Auto-setup team and sync on CLI login

18 Dec 13:54

Choose a tag to compare

0.5.5

fix: auto-setup team and sync on CLI login

v0.5.4 - Proper memory goal

14 Dec 21:08

Choose a tag to compare

add: postinstall message + bump version

- Update limitations page: remove "No semantic search yet", add hybrid search to "What's Ready"
- Update roadmap: remove v0.6 Smart Retrieval (shipped), renumber Agent Expansion to v0.6
- Update pricing: change "Cloud storage included" to "Shared team context"
- Update socials/footer: X link now points to @tryGrov
- Add postinstall message to drive npm users to dashboard
- Add Dashboard badge to README
- Add "Why Sync?" section to README

v0.5.3 - Hybrid Search

13 Dec 15:10
d12de82

Choose a tag to compare

What's New

🔍 Hybrid Search (RAG)

Memories now use semantic + lexical search for smarter context injection:

  • OpenAI embeddings (text-embedding-3-small) for semantic similarity
  • PostgreSQL full-text search for keyword matching
  • File boost (1.2x) for currently open files
  • Top 5 most relevant memories injected into context