C/OThe Companion OfficeAI companion rankings & research

Currency: USD

Tested August 2026 · 7 of 23 Lab tested · Methodology v0.1 · Updated 2026-08-24

Illustrative preview dataScores below are computed from a structurally complete staging dataset, not from published test findings about these products.

The best AI companion apps, tested and ranked

The answer right now

Rank 1Nomi81.7C/O Score

Top of the Overall ranking in this edition

Why it ranksVisit app ↗

What kind of companion each product is

No product carries a position yet: a coordinate needs a completed Basic Screen and a second reviewer's agreement.

See the complete map

  1. Rank 1Nomi81.7C/O Score

    Illustrative preview: in this staging dataset Nomi reads strongest on delayed recall and weakest on character fidelity under methodology v0.1.

    Illustrative record: character fidelity scored 74/100 in the staging scenario set.

    • delayed recall 94/1002026-08-15
    • character fidelity 74/1002026-08-15
    • Illustrative sample, paid-standard tier2026-08-15
    Why it ranksVisit app ↗
  2. Rank 2Character.AI79.4C/O Score

    Illustrative preview: in this staging dataset Character.AI reads strongest on character fidelity and weakest on memory controls under methodology v0.1.

    Illustrative record: memory controls scored 68/100 in the staging scenario set.

    • character fidelity 89/1002026-08-15
    • memory controls 68/1002026-08-15
    • Illustrative sample, paid-standard tier2026-08-15
    Why it ranksVisit app ↗
  3. Rank 3Kindroid79.2C/O Score

    Illustrative preview: in this staging dataset Kindroid reads strongest on delayed recall and weakest on character fidelity under methodology v0.1.

    Illustrative record: character fidelity scored 71/100 in the staging scenario set.

    • delayed recall 93/1002026-08-15
    • character fidelity 71/1002026-08-15
    • Illustrative sample, paid-standard tier2026-08-15
    Why it ranksVisit app ↗
  4. Rank 4Zeta76C/O Score

    Illustrative preview: in this staging dataset Zeta reads strongest on character fidelity and weakest on delayed recall under methodology v0.1.

    Illustrative record: delayed recall scored 65/100 in the staging scenario set.

    • character fidelity 87/1002026-08-15
    • delayed recall 65/1002026-08-15
    • Illustrative sample, paid-standard tier2026-08-15
    Why it ranksVisit app ↗
  5. Rank 5Replika75C/O Score

    Illustrative preview: in this staging dataset Replika reads strongest on conversation quality and weakest on character fidelity under methodology v0.1.

    Illustrative record: character fidelity scored 68/100 in the staging scenario set.

    • conversation quality 85/1002026-08-15
    • character fidelity 68/1002026-08-15
    • Illustrative sample, paid-standard tier2026-08-15
    Why it ranksVisit app ↗
  6. Rank 6Lovey Dovey74C/O Score

    Illustrative preview: in this staging dataset Lovey Dovey reads strongest on emotional response and weakest on character fidelity under methodology v0.1.

    Illustrative record: character fidelity scored 62/100 in the staging scenario set.

    • emotional response 87/1002026-08-15
    • character fidelity 62/1002026-08-15
    • Illustrative sample, paid-standard tier2026-08-15
    Why it ranksVisit app ↗
  7. Rank 7Talkie73.9C/O Score

    Illustrative preview: in this staging dataset Talkie reads strongest on character fidelity and weakest on memory controls under methodology v0.1.

    Illustrative record: memory controls scored 67/100 in the staging scenario set.

    • character fidelity 78/1002026-08-15
    • memory controls 67/1002026-08-15
    • Illustrative sample, paid-standard tier2026-08-15
    Why it ranksVisit app ↗

The full registry

23 products carry a full test bundle under the published method, and the highest of them are ranked above. The registry tracks 320 in all, each row carrying the evidence level it has earned and nothing it has not.

Open the full registry

Current research

Research