OpenMetadata
MCP servers became first-class governed assets in 1.13.0 — and 2.0 is now in release candidate.
A side-by-side editorial comparison of Hex and Mage — release velocity, themes, recent moves, and the top alternatives to consider.
| Feature | Hex | Mage |
|---|---|---|
| Sector | Analytics | Analytics |
| Velocity score | 6.3 | 0.0 |
| Sparks · 30d | 1 | 0 |
| Top themes | agents, cli, generative-apps, model-choice | data-pipelines, orchestration, cadence-decline, dependency-pinning |
| Last editorial update | 16h ago | 2h ago |
| Website | — | Visit → |
Hex pulls its agent out of the notebook and into a CLI and API
Hex ships weekly and the last two months read as one project: making the agent a surface in its own right rather than a feature living inside the notebook. The newest release exposes the agent through a CLI and an API, prints cell outputs in the terminal, and lists which resources each run touched. Generative apps got the same treatment slightly earlier, with editable app code, CSV downloads on every chart and table, and a signed embedding path.
Feature releases every two months in 2024; one bugfix release in the last twelve.
The release cadence has collapsed. Through 2024 Mage shipped roughly every two months with substantial features each time — memory management rework, dynamic blocks, new sources and destinations, Python 3.11 and 3.12 support. 2025 produced two releases. The most recent entry, 0.9.79 in January 2026, contains no feature section at all: it is dependency pinning, SQLAlchemy 2.0 compatibility, character escaping during code interpolation, and log file handle cleanup. Nothing has followed it in the six months since.
Hex ships weekly and the last two months read as one project: making the agent a surface in its own right rather than a feature living inside the notebook. The newest release exposes the agent through a CLI and an API, prints cell outputs in the terminal, and lists which resources each run touched. Generative apps got the same treatment slightly earlier, with editable app code, CSV downloads on every chart and table, and a signed embedding path.
Two things are converging: an agent that can be driven from anywhere, and generative apps as the artifact it produces. The model layer is deliberately swappable, with Claude Opus 5, Kimi K2.7, Fable 5 and an admin-set default all landing inside eight weeks, so Hex is not tying its product to one provider. Governance is arriving in step with the capability work, through spend limits, credit controls, enterprise role requests and IAM support, which is the sequence a tool follows when its buyers become platform teams.
The next move is likely unattended agent runs on a schedule or in CI, since the CLI and API remove the last reason a person has to be sitting in the notebook while it works.
The release cadence has collapsed. Through 2024 Mage shipped roughly every two months with substantial features each time — memory management rework, dynamic blocks, new sources and destinations, Python 3.11 and 3.12 support. 2025 produced two releases. The most recent entry, 0.9.79 in January 2026, contains no feature section at all: it is dependency pinning, SQLAlchemy 2.0 compatibility, character escaping during code interpolation, and log file handle cleanup. Nothing has followed it in the six months since.
The arc runs from expanding the product to keeping it compiling. The 2024 releases added capability — a canvas rework, multi-project support, streaming sinks, Kubernetes job parameters. The 2025 releases shifted toward integrations and CVE response, including a batch of path traversal fixes carrying assigned identifiers. The last release is entirely defensive, including vendoring croniter into the repository and locking scikit-learn to stop upstream changes from breaking builds. That is the profile of a codebase being kept viable rather than developed.
The entries give no basis for predicting the next release — a six-month gap after a dependency-only patch is the only signal available, and nothing here indicates whether the line is paused or finished.
Other Analytics products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either Hex or Mage.
MCP servers became first-class governed assets in 1.13.0 — and 2.0 is now in release candidate.
Every release in this window is columnstore work — compression is where TimescaleDB is spending
The streaming engine is stable and the API is being narrowed — Polars is clearing ground for a breaking release
Six releases, all patches — this window shows DuckDB's maintenance machine, not its roadmap
Two engines in one repo: the Python 1.x line tightens while Fusion 2.0 goes lakehouse-catalog native
Basedash turned its AI analyst into an API, then spent two weeks making it auditable
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
They serve adjacent needs but don't currently overlap on shipped themes. Hex is currently shipping more aggressively (velocity 6.3 vs 0.0), with 1 editorial sparks in the last 30 days against 0. See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. Hex is currently shipping more aggressively (velocity 6.3 vs 0.0), with 1 editorial sparks in the last 30 days against 0. For your specific use case, the alternatives sections above list other Analytics products to evaluate alongside.
Top Hex alternatives in Analytics are ranked by recent ship velocity. Browse the "Hex alternatives" section above for the current picks, or visit /alternatives/hex for the full list with editorial commentary on each.
Top Mage alternatives in Analytics are ranked by recent ship velocity. Browse the "Mage alternatives" section above for the current picks, or visit /alternatives/mage-ai for the full list with editorial commentary on each.