Rerun
Open-source data layer for Physical AI: log, query, transform, visualize, and train multimodal robotics data.
Rerun is the most comprehensive open-source data layer for Physical AI, uniquely combining visualization, query, transform, and training in one toolchain. The column-chunk storage, direct PyTorch dataloader, and recent additions like Viewer MCP and dataset review set it apart. However, Hub pricing requires a sales call, which may deter smaller teams. If you need a unified pipeline from sensor logs to model training, Rerun is a strong choice; for simpler ROS-only visualization, Foxglove might be simpler.
Verified 1d ago · liveness 81/100 · cite: rightaichoice.com/tools/rerun
- Robotics researchers and engineers building end-to-end learning pipelines
- Physical AI teams needing to visualize and debug multimodal sensor data
- Teams scaling from laptop experiments to cloud-based training
- Organizations wanting a unified data layer for collection, query, and training
- General-purpose BI or dashboarding (focused on Physical AI/time-series data)
- Teams that prefer fully-managed, no-ops platforms (Hub requires some setup)
- Users who only need simple charting and analytics without temporal data
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Rerun if you need a fully-managed, no-ops platform with transparent pricing and no sales calls, or if you only need simple visualization of ROS logs without training integration; in that case, consider lighter tools like Foxglove.
Hub pricing is contract-based with no public tiers, so you'll need a sales call and may hit minimum commitments as your team or data scales.
The SDK is free forever, which is ideal for individual developers, academics, and small teams. Hub pricing is tailored to team size and data scale, so it fits growing robotics companies but lacks transparency—contact sales for a quote. Compared to commercial alternatives like Weights & Biases, Rerun's free tier is more generous for the core data loop.
In short
Rerun — Open-source data layer for Physical AI: log, query, transform, visualize, and train multimodal robotics data. Best for Robotics researchers and engineers building end-to-end learning pipelines, Physical AI teams needing to visualize and debug multimodal sensor data, Teams scaling from laptop experiments to cloud-based training. Free to use.
What's new in Rerun
Checked yesterdayAcross the latest 4 updates: 1 feature update, 2 changelog entries and 1 news mention.
0.36.2
Patch release with bug fixes, including fixing opening external links.
0.36.1
Bug fixes for MCAP message recovery, dataframe API, and UI issues.
Rerun 0.36.0 released
Introduced a new data layer for physical AI, with migration guide included.
3D reconstruction is a data layer problem
Worked example using ARKitScenes scans converted into recordings that the viewer, dataframe API, and PyTorch dataloader read directly.
What people actually say about Rerun — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
45 mentions across 4 sources (Hacker News, App Store, GitHub, Lemmy) · researched Jul 3, 2026.
- +Unified data pipeline from logging to training in one tool.
- +Open-source SDK with permissive Apache-2.0/MIT license.
- +Efficient columnar storage for high-dimensional time-series data.
- +Interactive 2D/3D viewer ideal for robot sensor data.
- +Integrations with LeRobot and cuVSLAM add real-world credibility.
- −Over 1300 open GitHub issues signal reliability concerns.
- −Steep learning curve for beginners and non-robotics users.
- −Limited community discussion outside GitHub and niche forums.
- −Hub pricing not transparent; potential for unexpected costs.
- −New file format .rrd may not integrate with existing tools.
- • Hub pricing is not public—enterprises may face surprise costs.
- • Self-hosting Hub likely requires significant infrastructure.
Viability Score
How well maintained and how widely used is Rerun? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: August 2026
How we score →Key Features
- Log multimodal data with Python, C++, or Rust SDK
- Interactive viewer (desktop + web) with 2D, 3D, map, graph, tensor views
- SQL and dataframe queries over recordings and catalog
- Transform data with derived columns and schema evolution
- Train directly via PyTorch dataloader from .rrd files or Hub streams
- Store data as column-chunks in .rrd files for efficiency
- Declarative visualization framework with blueprints
- Byte-range indexing and retrieval from object storage (Hub)
- Team collaboration with shared recordings and link sharing (Hub)
- Open source SDK under Apache-2.0/MIT license
- Web viewer for browser-based visualization
- Chunk processing API for efficient data handling (0.32+)
- Dataset review for exploring recordings (0.32+)
- Viewer MCP server for AI agent integration (0.34+)
- HDF5 data import and video chunk reader with FFmpeg support (0.35+)
About Rerun
Rerun is an open-source SDK and commercial cloud platform purpose-built as the data layer for Physical AI, especially robotics. It unifies the entire data loop—logging, querying, transforming, visualizing, and training—on multimodal, multi-rate data using Python, C++, or Rust. The interactive viewer (desktop and web) supports 2D, 3D, map, graph, and tensor views, while SQL and dataframe queries provide direct access to recordings. The columnar .rrd file format stores data as column-chunks, enabling efficient queries and direct training via a PyTorch dataloader without export steps. Rerun Hub is the commercial backend that scales these same APIs to production, offering a managed catalog, byte-range indexing over object storage, streaming dataset mixes to GPUs, and team sharing with auth and SSO. It turns your object stores into a queryable, streamable foundation, and you can run transforms on the edge or close to the data. Unlike generic visualization tools like Foxglove or Weights & Biases, Rerun covers the entire data loop from logging to training, making it a unified alternative for robotics-centric workflows. The SDK is free forever under Apache-2.0/MIT, and Hub pricing is contract-based.
Behind the Verdict
Rerun stands out as a purpose-built data layer for Physical AI, covering the full spectrum from logging raw sensor data to training neural networks. Its strength lies in the deep integration of visualization, query, and training on a single data model. The columnar .rrd format and chunk processing enable efficient storage and retrieval, and the PyTorch dataloader allows training directly on recordings without export steps—a significant time-saver for robotics teams. The recent additions, such as the Viewer MCP server for AI agent integration, dataset review for exploring recordings, and HDF5 data import, show a commitment to evolving with the community's needs. The open-source SDK under Apache-2.0/MIT is genuinely free, with no asterisks, making it accessible to academics and hobbyists. However, the commercial Hub, while offering powerful features like byte-range indexing and GPU streaming, is not transparently priced—you must contact sales, which can be a hurdle for smaller teams or individual developers. Additionally, some features like MCAP support are experimental, and backwards compatibility of .rrd files is still maturing, so you may encounter breaking changes. For teams that need a unified, end-to-end data loop for robotics, Rerun is a compelling choice, but if you only need simple visualization of ROS logs, lighter tools like Foxglove might be more appropriate.
Researching Rerun? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Rerun actually fits — and what changes day-one when you adopt it.
Logging sensor data from a robot during a trial run
Outcome: Use the Python SDK to log images, LiDAR, and odometry, then open the recordings in the viewer to debug perception issues with synchronized 2D/3D views.
Building a training dataset from recorded demonstrations
Outcome: Query the catalog with SQL to filter relevant segments, then use the PyTorch dataloader to stream data mixes directly to GPUs, eliminating export pipelines.
Moving from local files to cloud-based storage for a growing team
Outcome: Adopt Rerun Hub to manage a catalog over S3, enable team sharing with SSO, and run transforms on the edge, all using the same SDK APIs.
Use Cases
- Visualize multimodal robot logs with synchronized 2D, 3D, and scalar time-series views.
- Query historical recordings using SQL or dataframe to debug perception failures.
- Stream training data mixes directly to GPUs from object storage without export steps.
- Transform raw sensor data with derived columns and share refined datasets across the team.
- End-to-end data loop for robotics: log, store, query, transform, train, and iterate.
Limitations
- Rerun's open-source SDK is free, but the commercial Hub is priced per-deployment with no public tiers, requiring contact with sales.
- The Hub runs in the cloud region of your choice, with your data in your own S3-compatible object storage, and self-hosting is only available for large teams with specific requirements.
- Some features, like MCAP support, are experimental, and backwards compatibility of .rrd files is still maturing.
- The pricing page indicates that Hub is managed by Rerun, with self-hosting generally not needed.
as of 2026-08-22
Verification history
We have re-verified Rerun 5 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Rerun tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Open Source SDK
$0
Ideal for
Individual developers, researchers, and small teams who want a free, self-hosted toolchain for logging, visualizing, and training on multimodal data without paying for infrastructure.
What this tier adds
This is the free entry point: full SDK access, local catalog, and community support on Discord and GitHub.
Commercial Hub
Contact us
Ideal for
Scaling robotics teams and enterprises that need a managed backend, team collaboration, SSO, and large-scale data streaming to GPUs without managing their own infrastructure.
What this tier adds
Adds managed catalog, byte-range indexing, link sharing, auth/SSO, and single-tenant isolation, with pricing based on team size and data scale.
Where the pricing makes sense
The company stage and team size where Rerun's pricing actually pencils out — and where peers do it cheaper.
The SDK is free forever, which is ideal for individual developers, academics, and small teams. Hub pricing is tailored to team size and data scale, so it fits growing robotics companies but lacks transparency—contact sales for a quote. Compared to commercial alternatives like Weights & Biases, Rerun's free tier is more generous for the core data loop.
Setup time & first value
How long it actually takes to get something useful out of Rerun — broken out by persona, not the marketing-page minute.
For Python users: install via pip install rerun-sdk and log your first recording within minutes. C++ and Rust users may need a bit more time for build setup. Exploring the viewer and dataframe queries takes a few hours. Setting up Hub for cloud-scale use requires a sales call and configuration, potentially days to weeks.
Switching to or from Rerun
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From ROS bags: Use the MCAP importer to convert your existing ROS bag files into .rrd recordings.
- →From Foxglove: Export your data as MCAP or HDF5 and import into Rerun for a more unified data loop.
- ↗To Foxglove: Export Rerun recordings to MCAP format for use in Foxglove if you only need visualization.
- ↗To Weights & Biases: If you prefer a commercial platform with more dashboarding features, you may need to export data to a standard format.
Integrations
Resources & Guides
Tutorials & Learning
Official links
Tools that pair well with Rerun
Common stack mates teams adopt alongside Rerun, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Rerun vs Spider Cloud
Spider Cloud and Rerun serve completely different domains: Spider Cloud excels at fast, cost-effective web crawling for AI agents and RAG pipelines, while Rerun is purpose-built for robotics teams logging and training on multimodal sensor data. Choose Spider Cloud if your need is real-time web data extraction; choose Rerun if you're building Physical AI systems.
Rerun vs Screenplayiq
Rerun vs Temporal Ai
Temporal AI is the choice for teams building reliable AI agents and complex business workflows that need guaranteed execution and recovery. Rerun excels for robotics engineers needing to log, visualize, and train on multimodal sensor data. Choose Temporal if your pain is crash recovery and orchestration; choose Rerun if your pain is debugging and scaling physical AI data.
Alternatives to Rerun
View allFrequently Asked Questions
Best-of guides
Used Rerun? Help shape our editorial sentiment research.