Vectorize @Vectorizeio

Makers of Hindsight. Agent memory that lets your agents learn over time. https://t.co/9U8TmRzNHf vectorize.io Boulder, CO Joined December 2023

Tweets

1K
Followers

1K
Following

329
Likes

787

Nicolò Boschi @nicoloboschi

3 days ago

Jev is a good model for reranking especially to improve precision over recall, being able to discard irrelevant results without a heuristic score filtering we are still thinking other ways to include it in the platform and experimenting with it btw this release has a bunch of other features, check them out!

Vectorize @Vectorizeio

3 days ago

Hindsight 0.10.1 is out. A new reranker that takes recall@1 from 0.800 to 0.950 while running 4x faster, plus a bank you can finally pick up and move: clone it, export it, rename it in place. Highlights below 🧵

3 2 22 2K 12

1 3 8 560 1

View Details

Jamie Connor @jamieconnor

17 hours ago

I just had a look at @Vectorizeio Hindsight in MCP mode for augmenting your LLM's with memory and it looked pretty impressive. Also works for large teams wanting collective memory. github.com/vectorize-io/h…

1 3 7 121 0

View Details

Nicolò Boschi @nicoloboschi

5 hours ago

0 1 7 627 0

View Details

Nicolò Boschi @nicoloboschi

5 hours ago

we're trending again on Github!

Trending GitHub Repositories @trending_repos

9 hours ago

Trending repository of the day 📈 hindsight Hindsight: Agent Memory That Learns Last 24h: 1,607 ⭐ Total: 26,931 ⭐️ github.com/vectorize-io/h…

0 4 74 8K 94

0 1 4 1K 4

View Details

Trending GitHub Repositories @trending_repos

9 hours ago

Trending repository of the day 📈 hindsight Hindsight: Agent Memory That Learns Last 24h: 1,607 ⭐ Total: 26,931 ⭐️ github.com/vectorize-io/h…

0 4 74 8K 94

View Details

rikkarth @rikkarth

12 hours ago

I really like Hindsight I genuinely feel since I've adopted this stack my model has become better and I don't feel like the need of repeating certain facts. This is what I'm using right now on a Pi4 w/ 4GBs RAM and 4 vCPU, a small brain always available anywhere I go. - Hindsight (Knowledge/Intelligence/Memory) - fastCRW (Context cleanup) - SearXNG (Reliable Web Search) - Graft (Project Knowledge Indexer) - rtk (tool calling compression/cleanup)

Nicolò Boschi @nicoloboschi

13 hours ago

what happens when you recall/search memories in Hindsight?

3 3 29 2K 31

2 4 9 587 6

View Details

Vectorize @Vectorizeio

8 hours ago

@prav1411 Great write up!

0 0 1 18 0

View Details

PK 🧠 @prav1411

7 days ago

I gave my AI memory so I wouldn't have to repeat myself. Then babysitting its memory became another project. 8,455 records later, I moved from Honcho to Hindsight. I wrote up what broke, what I kept, and why saved isn't the same as remembered. praveenks.com/notes/from-hon…

3 1 4 654 1

View Details

Nicolò Boschi @nicoloboschi

13 hours ago

what happens when you recall/search memories in Hindsight?

3 3 29 2K 31

View Details

Vectorize @Vectorizeio

19 hours ago

Cross-Encoder Reranking: The Last Stage of Agent Memory Recall. Retrieval is the part everyone talks about. Embeddings, BM25, graph traversal, time windows: pick your arms, fuse the rankings, ship it. But fusion doesn't tell you what's relevant. It tells you what several independent systems each thought was plausible, blended into one list. The stage that decides what your agent actually reads is the one that runs last, and it's the one that gets the least attention. This is a look inside that stage in Hindsight — the cross-encoder reranker. How it scores, why the secondary signals are multiplicative instead of additive, the subtle failure mode that turns a reranker into a date sort, and how to configure it so an unreachable model degrades instead of taking recall down with it. Know more: linkedin.com/pulse/cross-en…

5 0 5 401 2

View Details

Vectorize @Vectorizeio

a day ago

Migrating agent memory carries the old system's duplicate conclusions with it. Worse: imported facts arrive marked as already consolidated, so your bank never rebuilds the beliefs itself. Bring the source material. Leave the conclusions. x.com/i/article/2102…

0 0 2 242 0

View Details

Nicolò Boschi @nicoloboschi

2 days ago

what happens when you retain/write a document in Hindsight?

8 13 164 9K 218

View Details

Vectorize @Vectorizeio

2 days ago

Great Coding Agents Don't Just Reason. They Remember. Our New Plugin Makes That Standard. Hindsight 0.9.0 ships two things we've been building in parallel for months: Knowledge Pages — living documents your memory bank writes and maintains about itself — and a unified coding-agents plugin that brings long-term project memory to ten coding agents — Claude Code, Codex CLI, Cursor CLI, opencode, GitHub Copilot CLI, Cline, Kilo, Grok Build, Antigravity, and Devin — with one command and zero configuration. Neither looks like what we set out to build. This post is the story of how we got here, because the detours are the product: a benchmark we had to invent, an architecture of ours it killed, and a wiki that learned to heal itself. Know more: lnkd.in/gU9YvzhY

1 0 5 333 2

View Details

Vectorize @Vectorizeio

2 days ago

x.com/i/article/2102…

1 1 2 254 1

View Details

Vectorize @Vectorizeio

3 days ago

x.com/i/article/2102…

0 0 2 342 1

View Details

Vectorize @Vectorizeio

3 days ago

Plus faster retain on chunked documents, bank lists that don't crawl on big deployments, archive creation off the event loop, no more truncated Anthropic responses, and non-English content that keeps its own script. Full notes: hindsight.vectorize.io/blog/2026/09/2…

0 0 2 143 1

View Details

Vectorize @Vectorizeio

3 days ago

Two more that quietly save money: Automatic mental-model refreshes can run on their own cheaper model instead of the one your retains use. Memories and documents now filter by time on the server, so you stop pulling pages you were going to throw away.