Stop guessing what your AI infrastructure costs. Hivemind is one monthly price with a high usage ceiling. No usage math and no surprise bills. It pays for itself too: 1. Traces become skills 2. Agents stop repeating work 3. Token spend drops 33%. Make sure you have continual learning on for your AI agents.
Activeloop
Software Development
Mountain View, California 6,431 followers
Building Deeplake: the GPU-native, sandboxed Postgres for AI agents. → deeplake.ai
About us
Building Deeplake: the GPU-native, sandboxed Postgres for AI agents. Try out Deeplake today via deeplake.ai
- Website
-
https://deeplake.ai/?utm_source=linkedin&utm_medium=web&utm_campaign=linkedin_bio&utm_content=learnmore
External link for Activeloop
- Industry
- Software Development
- Company size
- 11-50 employees
- Headquarters
- Mountain View, California
- Type
- Privately Held
- Founded
- 2018
- Specialties
- Data Science, AI, Artificial Intelligence, Data pipelines, Cloud computing, Machine Learning, Computer Vision, Generative AI, Vector Search, LLMs, and Large Language Models
Employees at Activeloop
View 25 employees at Activeloop
or
Locations
-
Primary
Get directions
196 Castro St
Mountain View, California 94041, US
Updates
-
We have 3 new updates to Hivemind that are in service of our mission: making organizational agents more cost efficient and effective. 1. Proactive search experience for Claude Code + Cursor. After each user prompt, Hivemind automatically searches for relevant stored traces from the past and passes them to the agent as additional context. This improves reasoning and output. 2. Ability to share skills across teams and devices. Skill sharing is a simpler feature: it syncs the skills a user already has in their system through Deeplake, so everyone on the team can access them. Previously, we were more focused on generating and improving skills. This is about sharing existing skills across the team. 3. Ability to create and assign goals. Any teammate can create and assign goals to another that can persist across sessions until marked completed. Agents can automatically reason when the goal has been reached, and will notify the creator of the goal automatically. All of these new features work to level up your team, reduce errors and redundant work and lower AI token spend. Get set up with one command line install.
-
ScrapeGraphAI + Hivemind unifies intenral and external knowledge for coding agent.
Every day a new model came out and we are constantly switching between codex and Claude code. In this process we loose all the memory between them. Hivemind from Activeloop is solving this problem and create the best skill for improving your coding agent. So we decided to make a collaboration! Link for the article: https://lnkd.in/dtT84U2A
-
-
Hivemind helps coding agents get smarter with every team interaction, across all your agents, not just one. SkillOpt is what makes it real: your skills don't just accumulate, they get trained on your own traces and sharpened over time. The result is measurable. +19.1 points of accuracy in Claude Code, +24.8 in Codex, best or tied on all 52 setups tested. Your codebase becomes a graph-based knowledge base, helping your agent retrieve the right context beyond simple ranking.
Coding agents are getting longer context, better base models, more tools. But, none of it makes Friday's agent know what Monday's agent already figured out. That's the gap we built Hivemind to close. Hivemind is a continual learning layer for coding agents. It captures the traces from every agent your team runs (Claude Code, Codex, Cursor, OpenClaw, Hermes), turns the repeated patterns into reusable skills, and propagates those skills across all of them. An engineer's breakthrough on Monday is in everyone's agent context on Tuesday. Today we're adding the piece that was missing: the skills get self optimized. We've implemented SkillOpt, a text-space optimizer out of Microsoft and collaborators, directly into Hivemind. It improves each skill the way you'd tune a model, keeping only the edits that prove out on a held-out test. > The approach adds +19.1 points of accuracy inside Claude Code and +24.8 inside Codex, and wins or ties on all 52 setups tested. > It runs offline, so it costs nothing at inference. Your skills get sharper over time instead of bloating. > And all of it stores on your cloud storage. Your traces, your skills, your storage. For the engineering leaders here, that's the part that compounds: capability that survives onboarding, team changes, and turnover, instead of evaporating at the end of every session. One command and your team's agents share one brain.
-
We collaborated with AgentField.ai to deploy multi-agentic annotation system for physical AI. Here is Santosh Kumar Radha blogpost https://lnkd.in/ef9RCrkM
Multi-agent LLM systems have gone almost everywhere this year. Coding agents. Browser agents. Research agents. Customer support copilots. Robotics data curation hasn't really been on the list. That's the gap we worked on with the Activeloop (Deeplake) team. Roboscribe-AF is an open-source production ready annotation service that automates the work a human curator would otherwise do on a robot demonstration dataset — read the video, read the trajectory, segment the episode, check whether the two stories agree, and flag the low-confidence cases for review. The architecture splits cleanly: → Deeplake holds the versioned multimodal dataset. Video, states, actions, embeddings, branches. → AgentField runs the reasoning layer over each episode. 16 reasoners, 8 skills, an async workflow. The choice that matters most: the visual reasoner and the trajectory reasoner stay separate. A verifier compares them. When they disagree, we don't pick a winner — the disagreement gets written back to a Deeplake branch as a queryable field (consistency_score, human_review_recommended). The conflict becomes data the next training run can filter on. This is what we keep saying "AI backend" means. Not a chatbot pinned to a dataset viewer. Background reasoning attached to the pipeline itself — guided by schemas, branches, triggers, and review policies. Open source. Full write-up on the Deeplake blog: https://lnkd.in/ef9RCrkM
-
-
Frowg cooked with this one! They achieved 97%+ accuracy to provide organizational memory with Forge Cascade.
Frowg Systems, Inc.'s flagship product, Forge Cascade, a decentralized knowledge operating system, has integrated Activeloop's Deeplake and Hivemind. The results are impressive, read all the details in the article!
-
Focus on the work that matters. Let your agents powered by Hivemind handle the rest.
You shouldn’t have to chase your team every morning just to figure out what’s happening. But for most engineering managers, that’s still the routine: standups get skipped tickets go stale and writing a status update turns into detective work. What if your tools just told you? Imagine an AI layer that watches the real work happen across Claude Code, Codex, and OpenClaw, then surfaces: what changed who’s making progress and what’s stalled No nagging. No manual follow-ups. No extra process. You get your mornings back. Your engineers don’t feel watched. Everyone stays aligned. That’s the kind of workflow shift that actually matters.
-
Reading Jack Dorsey's "From Hierarchy to Intelligence", he gets a lot right. Namely: "Companies move fast or slow based on information flow." But framing it as a worker hierarchy problem is losing the plot. Look at where the actual work is moving: agents. Quick history: Email got messy. Slack fixed it. Then humans kept dropping balls anyway. Someone's offline, a thread dies, marketing has no idea what eng shipped, the handoff never happens. And now Slack itself is the slog. What if you could spend a fraction of the time in it? Meanwhile, your agents are in the pre-Slack era: • Your Claude Code agent has no clue what your coworker's OpenClaw agent decided yesterday. • Marketing's agent can't see what sales's agent promised the customer. • Product's agent has no idea what engineering's agent already shipped. Same company, same project, totally separate brains. The fastest workers on your team are stuck on the slowest part of your stack. Deeplake Hivemind fixes it. One install and your agents share memory across sessions, across teammates, across tools: Claude Code, OpenClaw, Codex, whatever. When one agent learns something, every agent on your team knows. No Slack pings. No status updates. No "wait, did you tell the VP?" Just shared context, flowing automatically. Slack was for humans. Hivemind is for the things actually doing the work now. Comment HIVEMIND and we'll DM you $100 in free credits. Run the experiment with your crew.
