diff --git a/CHANGELOG.md b/CHANGELOG.md
index a2841373..08f66118 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [3.4.1] - 2026-05-21
+
 ### Added
 - **ContinuousDiD methodology-review-tracker promotion.** Tracker row flipped **In Progress** → **Complete** with full Verified Components / Test Coverage / Corrections Made / Deviations / Outstanding Concerns structure mirroring the HAD precedent (PR #473). REGISTRY `## ContinuousDiD` gains a formal Deviations block consolidating the boundary-knots deviation from R `contdid` v0.1.0 (`range(dose)` vs `range(dvals)` — library avoids extrapolation), the `bspline_derivative` derivative-failure `UserWarning` (Phase 2 axis-C #12), the `+inf` → `0` never-treated recoding warning, and the zero-`first_treat`+nonzero-`dose` force-zeroing warning (both axis-E silent-coercion fixes) into a single AI-review-recognized labeled surface. R cross-language coverage for ContinuousDiD runs at relative tolerance across two surfaces: (a) **scalar parity with raw R `cont_did` / `pte_default`** at 1% on overall ATT for all 6 benchmarks and on overall ACRT for benchmarks 4-5 (benchmark 6 is event-study, scalar `overall_att` only); (b) **harmonized boundary-knot-normalized curve parity** with R-side ATT(d) / ACRT(d) reconstructed under `Boundary.knots = range(treated_doses)` (matching the library) on benchmarks 1-3 via the benchmark harness — `_run_r_contdid` does the R-side rebuild at `tests/test_methodology_continuous_did.py:333-367`, and `_compare_with_r` orchestrates the Python-vs-R comparison at `:395-459` — max ATT(d) at 1% and max ACRT(d) at 2%. NOT bit-exact (`atol=1e-8`) like HAD — the boundary-knots deviation precludes algorithmic bit-equality on aggregated dose-response curves. Surface (a) is direct raw-package parity; surface (b) is reconstructed-basis parity because raw `contdid` curves use `range(dvals)`. No source code changes, no new tests, no new docstrings — consolidation only against the existing 15 methodology tests (`tests/test_methodology_continuous_did.py`), 80 unit tests (`tests/test_continuous_did.py`), and `docs/methodology/continuous-did.md` theory note. `METHODOLOGY_REVIEW.md` ContinuousDiD row promoted **In Progress** → **Complete**.
 - **`SpilloverDiD(vcov_type="conley", survey_design=...)` integration via stratified-Conley sandwich on PSU totals (Wave E.2).** Lifts the Wave E.1 `NotImplementedError` (`spillover.py:2201` upfront, `two_stage.py:217` helper-level) and adds spatial-HAC + design-based variance for the previously deferred composition. **Documented synthesis** of Conley (1999) spatial-HAC × Gerber (2026, arXiv:2605.04124) Proposition 1 Binder TSL (the Wave E.1 foundation) × Wave D Gardner GMM first-stage uncertainty correction (Butts 2021 §3.1 + Gardner 2022 §4) applied to SpilloverDiD's ring-indicator stage-2 design. No reference software combines all three ingredients on a two-stage influence function. **Mechanical composition (panel-aware):** preserves the library's existing `conley_lag_cutoff = 0` semantic at `diff_diff.conley._compute_conley_meat` ("within-period spatial only — exclude cross-period spatial pairs") by looping over periods. For each period `t`, SpilloverDiD's per-obs Hájek-weighted Wave D IF `psi_i` is aggregated to per-period PSU totals `S_psu_t[g] = sum_{i in PSU g, time t} psi_i` (via `np.add.at`); per-PSU spatial centroids are panel-constant (mean of per-observation `conley_coords` within each PSU, vectorized `np.add.at` sums / `np.bincount` counts); for each stratum the within-stratum sandwich is `M_h_t = (1 - f_h) * n_h/(n_h-1) * sum_{j,k in PSUs_h} K(d(centroid_j, centroid_k) / conley_cutoff_km) * (S_psu_t[j] - S_bar_h_t)(S_psu_t[k] - S_bar_h_t)'`, where K is the Bartlett kernel (SpilloverDiD currently exposes Bartlett only and hardcodes it; the survey helper accepts `"uniform"` too but exposing that on the SpilloverDiD constructor is a separate follow-up) and `d` is haversine / euclidean / callable per `ConleyMetric`. Cross-stratum kernel weights are exactly zero by sampling design (strata are independence partitions). Total meat is `sum_t sum_h M_h_t`. Cross-period spatial pairs are excluded by construction — the per-period loop matches the library's panel Conley contract exactly. **Reduction semantics (load-bearing for tests):** the orchestrator's panel-aware meat equals `sum_t` of per-period within-stratum stratified-Conley sandwiches on per-period PSU totals (pinned at `tests/test_spillover.py::TestSpilloverDiDWaveE2ConleySurveyDesign::test_b_panel_aware_per_period_sum_invariant`); single stratum (H = 1, FPC = inf) reduces to `sum_t` plain Conley sandwich on per-period PSU totals (NOT on time-collapsed totals). **Implementation:** new `_compute_stratified_conley_meat_from_psu_scores` helper in `diff_diff/survey.py` (parallel to existing `_compute_stratified_meat_from_psu_scores` 3-tuple `(meat, variance_computed, legitimate_zero_count)` contract; per-stratum loop replaces the inner `centered.T @ centered` with `_compute_conley_meat(scores=centered, coords=psu_coords_h, ...)` in cross-sectional mode); new dispatch wrapper `_compute_stratified_conley_meat` in `diff_diff/two_stage.py` (parallel to existing `_compute_binder_tsl_meat`, performs per-obs Psi → PSU aggregation + centroid derivation + dispatch to survey helper, intentionally drops `cluster_ids` at the dispatch boundary — see Restrictions). `_compute_gmm_corrected_meat` conley branch extended with `if resolved_survey is not None` routing to the new wrapper; the `resolved_survey is None` branch is bit-identical to Wave D. **Singleton-stratum `lonely_psu="adjust"` parity:** the survey helper mirrors the Binder helper's `continue` to skip the FPC scale on singleton strata (with `n_h = 1` the scale `n_h / (n_h - 1)` would divide by zero); the degenerate one-PSU kernel `K = [[K(0)]] = [[1.0]]` reduces to `centered.T @ centered`, matching Binder's singleton-adjust output. **Saturated `df_survey = 0` NaN-fail:** mirrors Wave E.1 (`_compute_stratified_conley_meat` returns NaN meat with `UserWarning` template "Wave E.2 stratified-Conley sandwich: df_survey = 0..." so callers can `pytest.warns(UserWarning, match="Wave E.2 stratified-Conley")`). **Public surface restrictions:** replicate-weight variance (BRR / Fay / JK1 / JKn / SDR) raises `NotImplementedError` (inherits Wave E.1 gate; per-replicate full refit is separate follow-up scope); `cluster=<col> + survey_design.psu + vcov_type="conley"` coerces `cluster=<col>` to PSU per Wave E.1's warn-and-use-PSU pattern (the Conley cluster product kernel becomes a no-op after PSU aggregation, so `cluster_ids` is intentionally not threaded into the inner Conley kernel call — every PSU is its own cluster post-aggregation, which would zero all cross-PSU pairs); LinearRegression-side `vcov_type="conley" + survey_design=` gate at `diff_diff/linalg.py:2853` remains (separate Bertanha-Imbens 2014 weighted-Conley "Phase 5" roadmap, not Wave E); DiagnosticReport routing for `SpilloverDiDResults(vcov_type="conley", survey_design=)` requires `_APPLICABILITY` / `_PT_METHOD` registration (separate Wave F PR). **Tests:** new `TestSpilloverDiDWaveE2ConleySurveyDesign` and `TestSpilloverDiDWaveE2ConleySurveyDesignEventStudy` classes in `tests/test_spillover.py` (bit-identical no-survey fallback; panel-aware per-period sum invariant on the orchestrator + helper composition; hand-computation methodology anchor; single-stratum ≡ plain Conley on PSU totals; cross-stratum independence as a unit test on the survey helper with interleaved cross-stratum centroids; Binder vs Conley singleton-adjust FPC skip parity; lonely-PSU sensitivity across three modes; FPC large ≡ no-FPC and FPC = n_h zeros stratum; saturated NaN-fail with `pytest.warns(match="Wave E.2 stratified-Conley")`; replicate-weight + non-pweight rejections; cluster warn-and-use-PSU; fit idempotency; `finite_mask` survey-array subsetting; no-PSU coverage — weights-only `SurveyDesign(weights=...)`, strata-only `SurveyDesign(weights=..., strata=...)`, and a per-period re-index unit invariant pinning that no cross-period spatial pairs leak into the meat on implicit-PSU layouts; event-study path on both `is_staggered=True`/`False` branches per `feedback_cohort_loop_trigger_cache_both_branches`; drift goldens at `rtol=1e-12 / atol=1e-14`). The pre-existing `tests/test_spillover.py::test_fit_conley_plus_survey_design_not_implemented` Wave E.1-era gate-assertion test is removed (replaced by the positive-path tests above). Wave E.1 entry's "Public surface restrictions" bullet updated to past-tense the conley+survey gate reference.
@@ -1490,6 +1492,7 @@ for the full feature history leading to this release.
 [2.1.2]: https://github.com/igerber/diff-diff/compare/v2.1.1...v2.1.2
 [2.1.1]: https://github.com/igerber/diff-diff/compare/v2.1.0...v2.1.1
 [2.1.0]: https://github.com/igerber/diff-diff/compare/v2.0.3...v2.1.0
+[3.4.1]: https://github.com/igerber/diff-diff/compare/v3.4.0...v3.4.1
 [3.4.0]: https://github.com/igerber/diff-diff/compare/v3.3.3...v3.4.0
 [3.3.3]: https://github.com/igerber/diff-diff/compare/v3.3.2...v3.3.3
 [3.3.2]: https://github.com/igerber/diff-diff/compare/v3.3.1...v3.3.2
diff --git a/CITATION.cff b/CITATION.cff
index 06075792..03670f35 100644
--- a/CITATION.cff
+++ b/CITATION.cff
@@ -7,8 +7,8 @@ authors:
     family-names: Gerber
     orcid: "https://orcid.org/0009-0009-3275-5591"
 license: MIT
-version: "3.4.0"
-date-released: "2026-05-19"
+version: "3.4.1"
+date-released: "2026-05-21"
 doi: "10.5281/zenodo.19646175"
 url: "https://github.com/igerber/diff-diff"
 repository-code: "https://github.com/igerber/diff-diff"
diff --git a/diff_diff/__init__.py b/diff_diff/__init__.py
index aee77654..8d278092 100644
--- a/diff_diff/__init__.py
+++ b/diff_diff/__init__.py
@@ -293,7 +293,7 @@
 DCDH = ChaisemartinDHaultfoeuille
 HAD = HeterogeneousAdoptionDiD
 
-__version__ = "3.4.0"
+__version__ = "3.4.1"
 __all__ = [
     # Estimators
     "DifferenceInDifferences",
diff --git a/diff_diff/guides/llms-full.txt b/diff_diff/guides/llms-full.txt
index 8bfbc55a..4dc815d8 100644
--- a/diff_diff/guides/llms-full.txt
+++ b/diff_diff/guides/llms-full.txt
@@ -2,7 +2,7 @@
 
 > A Python library for Difference-in-Differences causal inference analysis. Provides sklearn-like estimators with statsmodels-style output for econometric analysis.
 
-- Version: 3.4.0
+- Version: 3.4.1
 - Repository: https://github.com/igerber/diff-diff
 - License: MIT
 - Dependencies: numpy, pandas, scipy (no statsmodels dependency)
diff --git a/pyproject.toml b/pyproject.toml
index 7455c91e..35eca976 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "maturin"
 
 [project]
 name = "diff-diff"
-version = "3.4.0"
+version = "3.4.1"
 description = "Difference-in-Differences causal inference with sklearn-like API. Callaway-Sant'Anna, Synthetic DiD, Honest DiD, event studies, parallel trends."
 readme = "README.md"
 license = "MIT"
diff --git a/rust/Cargo.toml b/rust/Cargo.toml
index 734780a0..007025c6 100644
--- a/rust/Cargo.toml
+++ b/rust/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "diff_diff_rust"
-version = "3.4.0"
+version = "3.4.1"
 edition = "2021"
 rust-version = "1.85"
 description = "Rust backend for diff-diff DiD library"