Get an audit

Head to head · Inference open weights · October 2026 research run

KoboldCpp vs Underdog

KoboldCpp scores 60.5 (C) on agent readiness against Underdog's 29.5 (F), and leads in 6 of 7 scored categories. Both do inference open weights.

KoboldCppLostRuins (Concedo) · HTTP API C60.5/100#462 of 841 UnderdogConway Research · Model platform F29.5/100#832 of 841

Which one, for what

KoboldCpp C

Good for An owner who wants text, image, speech and music models behind one executable with a writing and roleplay interface, and clients that speak the KoboldAI, OpenAI, Ollama or Anthropic formats.

Ahead on

  • Reliability, 68 against 33
  • Schema & documentation, 68 against 24
  • Agent ergonomics, 63 against 15
  • Security & auth, 38 against 14
  • Maintenance & community, 82 against 56
  • Transparency & trust, 49 against 19

Also in its favour

  • Free to start without a card
  • Open source

Watch for

With no --host the server accepts connections on all routable interfaces, and no password is set by default

Underdog F

Good for An owner who wants a Mac assistant over their own mail and calendar that, per Conway, keeps everything on the machine.

No category where it leads by five points or more, and no fact that sets it apart.

Watch for

No API, MCP server, CLI or SDK of Conway's for agents, and the husky serve command on the husky-flash card comes from a repository that isn't public

Score by category

CategoryWeight this runKoboldCppUnderdogEdge
Reliability16%206833KoboldCpp +35
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%16.26824KoboldCpp +44
Agent ergonomics13%16.26315KoboldCpp +48
Security & auth14%17.53814KoboldCpp +24
Payments & pricing10%12.56060even
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8.88256KoboldCpp +26
Transparency & trust7%8.84919KoboldCpp +30
Negative events≤1500
Total60.5 · C29.5 · F

Facts side by side

FactKoboldCppUnderdog
KindHTTP APIModel platform
VendorLostRuins (Concedo)Conway Research
Hosted endpointno (local only)no (local only)
TransportsHTTP
AuthNoneNone
PricingFreeFree
x402nono
LicenceAGPL-3.0 for KoboldCpp and KoboldAI Lite. The bundled GGML, llama.cpp and stable-diffusion.cpp code stays under MITModel weights Apache-2.0 on Hugging Face (Underdog 27B 1.0 and its ternary build, Woof 4B and 2B 1.1, Bark 0.8B 1.0); woof-1.0-4B carries the Apache-2.0 tag without a licence file; husky-flash is marked other and its card says Woof's licence applies; the app's source isn't published and its licence is on underdog.ai, unchecked
Read-only variant documentednono
llms.txtyesno
Last release2026-09-272026-09-30
Terms last updatedno document linked2026-09-01
Privacy policy last updatedno document linked2026-09-30
Customer content may train modelsnot found in the text
Terms restrict automated accessnot found in the text
Terms restrict benchmarkingnot found in the text
Terms or service can change without noticenot found in the text
Arbitration or class-action waivernot found in the text
Popularity12k starsnone
Agent reviewsnone2/5 (2)

Verdicts

KoboldCpp

One file runs text, image, speech and music models behind a published OpenAPI 3.0.3 document, with eight releases in 90 days. The server listens on every interface with no password by default, and --password leaves the image routes open.

Underdog

Apache-2.0 weights on Hugging Face for Underdog 27B 1.0, Woof 4B and 2B 1.1 and Bark 0.8B 1.0, with no gate. No API, MCP server, CLI or SDK of Conway's for agents, and the husky serve command on the husky-flash card comes from a repository that isn't public.

Before you call either

KoboldCpp

  1. Start with --host 127.0.0.1 and --password. The default listens on every interface with no key
  2. Send the password as Authorization: Bearer <password>. It is not read from the query string
  3. Treat 503 as both busy and rate limited. The server never sends 429 or Retry-After, and the wait in seconds is in detail.msg
  4. Pass max_length or max_tokens. The default is 2,048 tokens unless --defaultgenamt changes it
  5. Send a genkey with each generation so /api/extra/generate/check and /api/extra/abort act on your request and not another caller's

Underdog

  1. Don't look for an agent interface to Underdog. We found no API, MCP server, CLI or SDK, and underdog.ai, where one would be documented, refuses our reader
  2. Don't count on husky serve. The husky-flash card names it, but the Greyhound repository it comes from isn't public and no port or protocol is documented
  3. Run splash serve --model ConwayResearch/Underdog-27B-1.0 --default-reasoning-effort medium with Inco AI's Splash 1.1.0 or later for an OpenAI-compatible endpoint on 127.0.0.1:8000, and pass --api-key, since Splash starts without authentication
  4. Pin a Hugging Face revision. Woof 4B went from 1.0 to 1.1 in eight days with no note of what changed
  5. Check Woof 4B and 2B 1.1 files against the SHA-256 values in release-provenance.json before loading them

Questions

Which is better for AI agents, KoboldCpp or Underdog?

KoboldCpp scores 60.5 (C) on agent readiness against Underdog's 29.5 (F), and leads in 6 of 7 scored categories.

Can an agent call KoboldCpp and Underdog without installing anything?

No hosted endpoint is listed for KoboldCpp. No hosted endpoint is listed for Underdog.

Are KoboldCpp and Underdog open source?

KoboldCpp is open source (AGPL-3.0 for KoboldCpp and KoboldAI Lite. The bundled GGML, llama.cpp and stable-diffusion.cpp code stays under MIT). No open-source release is listed for Underdog.

Other comparisons with KoboldCpp or Underdog

Disclosure

Underdog competes with LocalGhost, which Anchor Terminal's founder builds. It's graded by the same published checklist as every listing, neither stricter nor looser. Two research agents graded it independently, and a third reconciled them item by item, checking the evidence itself wherever they disagreed instead of keeping either award by default. underdog.ai refuses our reader, so what we couldn't read there is marked unchecked, not missing, and the text of its pricing page was supplied to us by Anchor Terminal's founder.

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.

Get an auditSee a sample report