Chart-ready data
Visual Data
Data tables for the MCP Memory Server Sector Radar. All figures are frozen at the 2026-05-10 evidence snapshot and follow the SC-013 display standard: denominator, source, and reader meaning must stay visible beside public metrics.
Display Contract
Reference implementation update for Operations/SECTOR-RADAR-DISPLAY-STANDARD.md. MCP Memory Servers metrics use either the in-cluster corpus of N=20 repositories or the full corpus of N=32 repositories. Each section below names the sample, source artifact, and reader meaning.
| Display Area | Denominator / Sample | Primary Source | Reader Meaning |
|---|---|---|---|
| Cluster composition | In-cluster, N=20; full corpus, N=32 | Visual Data Charts 1-2; calibration role assignments | Shows sector shape inside the retained cluster and the wider discovery boundary. |
| Health metrics | In-cluster, N=20 and full corpus, N=32 | Visual Data Chart 4; calibration score matrix | Compares convergence, differentiation, viability, scaffold rate, integrity, and leakage. |
| Centroid traits | In-cluster, N=20 | Visual Data Chart 5; centroid artifact | Shows which traits define the default MCP memory-server pattern. |
| Score scatter tables | Displayed examples from the full corpus, N=32 | Visual Data Charts 3 and 6; calibration table | Shows how sector position, differentiation, human engineering depth, and adoption relate. |
Cluster Composition
Denominator: in-cluster, N=20, and full corpus, N=32. Source: Reports/VISUAL-DATA.md Charts 1-2 and Manual-audit/CALIBRATION-TABLE-RADAR.md. Reader meaning: role share shows how many projects sit inside the sector default, how many are distinct, and how many appeared because adjacent infrastructure shares the discovery surface.
In-Cluster Roles
Full Corpus Roles
Health Metrics
These metrics compare the retained MCP memory-server cluster with the broader discovery corpus. Category Leakage here is discovery-boundary leakage: 8 of 32 full-corpus rows are Adjacent Infrastructure projects that appeared in the wider sector search space.
| Metric | Value | Scale | Denominator / Sample | Source | Reader Meaning |
|---|---|---|---|---|---|
| Corpus denominator | 20 | N | In-cluster, N=20 | Visual Data Source-of-Truth Table | Retained repositories used for centroid and in-cluster health claims. |
| Corpus denominator | 32 | N | Full corpus, N=32 | Visual Data Source-of-Truth Table | Wider discovery corpus including adjacent, reference, self, and edge-case rows. |
| Convergence Index | 4.35 | 0-5 | In-cluster, N=20 | Visual Data Chart 4 | High convergence around the MCP memory-server centroid. |
| Convergence Index | 3.5 | 0-5 | Full corpus, N=32 | Visual Data Chart 4 | Lower corpus-wide convergence once adjacent infrastructure is included. |
| Differentiation Index | 3.1 | 0-5 | In-cluster, N=20 | Visual Data Chart 4 | Moderate differentiation inside the retained cluster. |
| Differentiation Index | 3.4 | 0-5 | Full corpus, N=32 | Visual Data Chart 4 | Higher differentiation when adjacent infrastructure is included. |
| Viability Rate | 60% | % | In-cluster, N=20 | Visual Data Chart 4 | Share of retained projects showing enough sustained engineering to look viable. |
| Viability Rate | 66% | % | Full corpus, N=32 | Visual Data Chart 4 | Corpus-wide viability rises because adjacent infrastructure includes mature projects. |
| Scaffold Rate | 25% | % | In-cluster, N=20 | Visual Data Chart 4 | One in four retained MCP memory servers is structurally present but implementation-light. |
| Scaffold Rate | 16% | % | Full corpus, N=32 | Visual Data Chart 4 | Scaffold share drops across the wider corpus. |
| Integrity Index | 3.8 | 0-5 | In-cluster, N=20 | Visual Data Chart 4 | Claims generally map to implementation evidence inside the retained cluster. |
| Integrity Index | 3.9 | 0-5 | Full corpus, N=32 | Visual Data Chart 4 | Corpus-wide claim/implementation integrity remains similar. |
| Category Leakage | 25% | % | Full corpus, N=32 | Visual Data Chart 4 | Discovery-boundary leakage: 8/32 rows are adjacent infrastructure rather than retained MCP memory servers. |
| Adoption Concentration | 15% | % | In-cluster, N=20 | Visual Data Chart 4 | Only a small retained subset reaches high adoption depth. |
Centroid Trait Prevalence
Denominator: in-cluster, N=20. Source: Reports/VISUAL-DATA.md Chart 5 and Manual-audit/CENTROID-MCP-MEMORY-32.md. Reader meaning: prevalence shows which evidence-backed traits define the MCP memory-server default pattern.
Score Matrix Context
Denominator: full corpus, N=32, with in-cluster rows marked separately where relevant. Source: Manual-audit/CALIBRATION-TABLE-RADAR.md and the visual-data tables on this page. Reader meaning: the score matrix explains how public role labels are derived without turning a repo row into a universal quality verdict.
| Score Key | Scale | Question Answered | Primary Source |
|---|---|---|---|
| CM | 0-5 | How closely the repo matches the MCP memory-server centroid. | Centroid match scoring table and calibration rows. |
| DF | 0-5 | How much observable differentiation appears beyond the default pattern. | Calibration rows and source notes. |
| HE | 0-5 | How much sustained engineering evidence is visible. | Calibration rows and audit traces. |
| AD | 0-5 | How much adoption signal exists at the snapshot date. | Calibration rows and public repository metadata. |
| CI | 0-5 | Whether claims match implementation evidence. | Calibration rows and source trace notes. |
Centroid Match vs Differentiation
Sample: displayed examples from the full corpus, N=32. Source: Reports/VISUAL-DATA.md Chart 3 and calibration table. Reader meaning: CM means Centroid Match and DF means Differentiation; together they show whether a repo is close to the default sector pattern, distinct within it, or adjacent to it.
| Repo | CM | DF | Role | Category |
|---|---|---|---|---|
| claude-mem | 5 | 4 | Category Leader | In-Cluster |
| mcp-memory-service | 4 | 5 | Category Leader | In-Cluster |
| MemPalace | 5 | 5 | Differentiated Niche | In-Cluster |
| memory-bank-mcp | 3 | 1 | Baseline Implementation | In-Cluster |
| mem0-mcp | 4 | 0 | Inactive Scaffold | In-Cluster |
| Mem0 | 1 | 5 | Adjacent Infrastructure | Adjacent |
| Graphiti | 0 | 4 | Adjacent Infrastructure | Adjacent |
| Rekindle | 5 | 5 | Differentiated Niche | Reference/Self |
| mcp-server (official) | 3 | 2 | Reference Implementation | Reference |
| claude-memory-compiler | 2 | 3 | Edge Case | Edge Case |
Adoption Depth vs Human-Edge
Sample: displayed examples from the full corpus, N=32. Source: Reports/VISUAL-DATA.md Chart 6 and calibration table. Reader meaning: HE means Human-Edge and AD means Adoption Depth; this view separates sustained engineering from public reach.
| Repo | HE | AD | Role | Category |
|---|---|---|---|---|
| claude-mem | 4 | 5 | Category Leader | In-Cluster |
| mcp-memory-service | 5 | 5 | Category Leader | In-Cluster |
| ByteRover | 5 | 2 | Differentiated Niche | In-Cluster |
| memory-bank-mcp | 2 | 5 | Baseline Implementation | In-Cluster |
| persistent-context-mcp | 0 | 0 | Inactive Scaffold | In-Cluster |
| Mem0 | 5 | 5 | Adjacent Infrastructure | Adjacent |
| Graphiti | 5 | 5 | Adjacent Infrastructure | Adjacent |
| Rekindle | 5 | 1 | Differentiated Niche | Reference/Self |
| mcp-server (official) | 4 | 5 | Reference Implementation | Reference |
| supermemory-mcp | 2 | 3 | Baseline Implementation | Edge Case |
Source Artifacts
All source data for these tables is published alongside the report. Found a stale repo, wrong source trace, or boundary classification issue? Send a correction to hello@diversum.dev with the repo name, displayed value, and source that should replace it.
- Centroid definition — 9-trait default pattern and denominator protocol
- Calibration table — all 32 repositories scored with derivation traces
- Methodology — Sector Radar method, public language, and correction process
| Published Artifact | Source Role | Visible Metric Families |
|---|---|---|
| Manual-audit/CENTROID-MCP-MEMORY-32.md | Centroid traits and prevalence counts | Centroid trait prevalence and in-cluster denominator. |
| Manual-audit/CALIBRATION-TABLE-RADAR.md | Per-repo scores, evidence notes, cluster roles, and category labels | Role composition, score examples, health metrics, and leakage. |
| Reports/VISUAL-DATA.md | Chart-ready aggregations derived from calibration | All displayed chart and table aggregations on this page. |
| Reports/SECTOR-RADAR-MCP-MEMORY-SERVERS.md | Public narrative report | Reader framing, sector interpretation, limitations, and methodology references. |
| Operations/Sprint-Contracts/SC-010.md | Original MCP Memory Servers report-generation scope | Report package acceptance context. |
| Operations/SECTOR-RADAR-DISPLAY-STANDARD.md | SC-013 display contract | Denominator, source trace, public label, role color, and correction-path requirements. |