DX alternatives for teams that want measured throughput
The shortlist is Navigara, Jellyfish, LinearB, Swarmia, Faros AI, and Waydev. All six read repository and delivery data, so each produces a figure without asking anyone to fill in a form. DX measures something those six cannot see at all, which is what developers experience while doing the work, so the useful comparison is about which question you have been asked rather than which tool is better.
What DX measures, and why it is hard to substitute
DX is an Atlassian company. The acquisition was announced on 18 September 2025 and completed on 10 November 2025. The $1B figure comes from DX’s own announcement and press coverage rather than from Atlassian, which did not disclose terms. DX still operates under its own brand and has since published a commitment to data neutrality.
Its headline number is DXI, which DX describes as a composite score of 14 key drivers of engineering efficiency, scored out of 100. The input is largely survey responses. That method reads friction: a build that takes 40 minutes, documentation nobody trusts, an on-call rotation that eats a sprint, a review queue where changes sit for three days. None of that appears in a commit. A repository-based instrument can measure the consequence of slow builds and never identify the cause.
DX also holds a clear published position on individual use: “Never tie throughput metrics to individual performance.” Among tools in this category, that stance is not universal, and it is worth crediting.
Why teams look elsewhere anyway
Three specific reasons, and none of them is that surveys are unserious.
Response rate is a dependency. A sentiment composite needs people to answer, and to answer candidly. Both depend on trust that already exists. A team that has just been through a layoff cycle will produce a DXI that measures the layoff.
A survey cannot produce a before-and-after for a tool adoption that has already happened. If AI assistants were introduced into the workflow 18 months ago and nobody ran a baseline survey first, that baseline does not exist and cannot be reconstructed. Repository history can be scored backward.
Finance does not act on sentiment. A CFO asked to renew a six-figure tooling line wants a measured change, and a driver score is a weaker instrument in that specific conversation than in the one about where to invest platform effort.
The six alternatives
Navigara scores each merged change by what it was, a complex refactor, a performance fix, a feature unblock, a dependency bump, and compares the result to that team’s own pre-AI window. Measurement runs at team and repository level with no per-developer throughput view. Pricing is published: $7 per developer per month up to 30 developers, $30 on Pro, a free 14-day Explore tier, $4,500 for a one-off private benchmark. It does not publish a capitalization feature, so a finance team needing capitalization should look at the others here.
Jellyfish integrates Git and planning data into roster and payroll, allowing it to express engineering work in dollars. That makes it the strongest of the six for R&D capitalization and board investment-mix reporting. It reports at the individual level and has published its own guidance against using metrics that way. No published pricing.
LinearB reads Git and pull request telemetry alongside trackers and CI/CD, and benchmarks cycle time, DORA, and pull request size against 8.1M+ pull requests. That comparison set is a genuine asset. Published pricing is $29 and $59 per user per month, and the tiers are built for teams of 50+ and 100+ developers, which rules it out below about 50 engineers. Its AI spend attribution is not published.
Swarmia is the closest in spirit to DX while still reading delivery data, because it combines Git, trackers, CI/CD, surveys and 22 HRIS and payroll systems. Its headline is Investment Balance, allocation by time and money, and it publishes no composite score by choice. Its stance is explicit: “Say bye to leaderboards and stack ranking.” Pricing is published: free for 9 developers, then $45 and $55; capitalization at $18 and AI cost at $5 per developer per month. It is the only platform here claiming a SOC 1-audited capitalization approach.
Faros AI claims to have 60+ sources and publishes the strongest AI spend attribution in the set, including a Token Attribution Ledger and cost-per-verified-outcome reporting. It no longer publishes capitalization pages, and the old capitalization URLs no longer resolve to capitalization content, so that feature’s status is unclear. No published pricing.
Waydev reads Git, pull request telemetry, and Jira, and reports a composite Impact score derived partly from lines of code at the organization, team, and developer levels. Published pricing is $29 and $49 per active contributor per month, billed annually. Its only published caution on metric use is “Don’t use metrics to micromanage.”
Side by side
| Primary data | Headline number | AI spend attribution | Per-individual reporting | Published pricing | |
|---|---|---|---|---|---|
| DX | Survey-heavy mix | DXI, 14 drivers | Token-level spend by team, contributor and tool | Published stance against it | None |
| Navigara | Git repository history | Throughput against own pre-AI baseline | Cost measured against throughput change | No such view | Yes |
| Jellyfish | Git, planning, roster and payroll | Investment allocation in R&D dollars | Token level, by tool, team or initiative | Yes, with guidance against | None |
| LinearB | Git, PR telemetry, trackers, CI/CD | Cycle time and DORA, benchmarked | Not published | Team-first, role-based access | Yes |
| Swarmia | Git, trackers, CI/CD, surveys, HRIS | Investment Balance | Cost by person per tool | Published stance against it | Yes |
| Faros AI | 60+ sources claimed | Cost per verified outcome | Token Attribution Ledger | No stated position | None |
| Waydev | Git, PR telemetry, Jira | Impact composite | Tool, team and seat | Yes | Yes |
When to keep DX
Keep it when the question you are being asked is why delivery is slow rather than whether it changed.
A VP told to fix a six-week lead time needs to know which stage is absorbing the weeks and why, and the answer is frequently something no repository records: a staging environment that breaks twice a week, a release process requiring three manual approvals, a review culture where nobody wants to be the one to block. DXI surfaces that. A throughput score reports its cost.
Many orgs with roughly 150 engineers or more should run both, and Swarmia’s own design makes the point by including surveys alongside delivery data. The two instruments are not competing for the same slot on the same slide.
If the question in front of you is whether the throughput number actually moved, connect a repository and the pre-AI window scores in the first pass.
How to choose
Three questions settle it, in this order.
What have you been asked for? A renewal decision or a board number points to measured delivery data. A fix-the-friction mandate points to DX.
Does the baseline you need still exist? If the tool adoption you want to evaluate has already happened, only repository history can be scored backward.
Who will see a per-person view? Four of the seven options here can produce one. Decide that before the rollout, because the announcement sets the tone for how the tool is received, and you only get one.
Talk to us if you want a measured baseline comparison alongside whatever survey instrument you already run.
Frequently asked questions
- Is DX still independent after the Atlassian acquisition?
- DX operates under its own brand and has published a commitment to data neutrality since the acquisition, which was announced on 18 September 2025 and completed on 10 November 2025. Treat it as an Atlassian company when weighing long-term product direction.
- Can DX measure the ROI of AI coding tools?
- DX publishes an AI cost management report covering token-level spend by team, contributor, and tool. The sentiment side indicates whether developers feel the tools help, which is useful and a different claim from a measured change in throughput.
- What is the closest DX alternative?
- Swarmia, because it combines surveys with delivery data and takes a comparable published position against leaderboards. For repository-measured throughput against a pre-AI baseline specifically, Navigara.
- Do I have to choose between surveys and repository data?
- No, and above about 150 engineers, most orgs should run both. They answer different questions, and the combined view is stronger than either alone.
- Which of these publish pricing?
- Navigara, LinearB, Swarmia, and Waydev. DX, Jellyfish, and Faros AI publish none, so expect a sales cycle before you see a number.
- Does Navigara replace DXI?
- No. There is no sentiment measurement in Navigara at all, so anything DXI tells you about developer experience has no equivalent there.
