# One chat interface over local models for a household or a team

> One chat interface over local models for a household or a team: 0 companies fully do it, 3 with a limit, 133 carry the tag. Every listing in Local AI with the reason and a source.

- Canonical: https://www.anchorterminal.com/niches/local-ai.household-or-team-chat/
- Markdown: https://www.anchorterminal.com/niches/local-ai.household-or-team-chat/index.md (~1,550 tokens)
- Slim: https://www.anchorterminal.com/niches/local-ai.household-or-team-chat/index.min.md (~280 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/niches/local-ai.household-or-team-chat/index.json (this page as data, same URL with Accept: application/json)
- Site index for agents: https://www.anchorterminal.com/llms.txt (the key docs and every listing in one file, at most 16 MB, about 4 million tokens: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- MCP: https://www.anchorterminal.com/mcp (the whole catalogue as tools, no key)
- DNS: dig +short TXT <slug>.grade.anchorterminal.com (the grade as a DNS TXT record, for a sandbox with no HTTP where names still resolve; also top., best.<category>. and <a>.vs.<b>. under the zone, and the same under grade.letme.dev)
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)
- Updated: 2026-11-15

No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.

| Companies, full fit | With a limit | Carry the tag | Share of the category's work | First-hand accounts |
| --- | --- | --- | --- | --- |
| 0 | 3 | 133 | 13 | 9 |

## What counts as done

Two or more users log in, chat with the same local model concurrently, and an administrator can see per-user access; the model answers with the attached knowledge where asked.

Measured, for our own tests: a network monitor sees no prompt text leave the machine.

How agents ask for it: set up a private ChatGPT for my family; give staff a chat over our internal documents on our own server; a chat UI for the model running on my home server; deploy private LLM chat for a few hundred users

## Does the job, with a limit

It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.

- **AnythingLLM (Mintplex Labs, Inc.)**: Docker build has multi-user mode, invites, admin and rule-based access to workspaces. API keys are admin-equivalent with no scopes. One account rejected it for the install size and setup.
- **LocalAI (Ettore Di Giacinto and the LocalAI team)**: The README's 4.1.0 notes announce a multi-user platform with OIDC, API keys and per-user quotas, and saved chats with agents. The multi-user docs page was not read, so accounts and groups rest on release notes.
- **Open WebUI (Open WebUI Inc.)**: Roles, groups, permissions and per-user API keys (admin must enable them, no expiry). Deployments for 700 to 800 users exist. Accounts report persona prompts drifting after a few turns and configuration effort.

## Couldn't check

The docs were closed to us, or silent. These carry no mark either way.

- **Core (ZMJ, Inc., doing business as Ghost, 325 9th Street, San Francisco, CA 94103, United States)**: The FAQ says desktop and phone clients connect; no accounts, per-user access or concurrency is documented.
- **Khoj (Khoj Inc.)**: Setup docs cover anonymous mode and per-user kk- API keys, but the authentication and multi-user page was not read, so per-user access could not be confirmed.

- not this job: Berth (Trevin Tindall (T-Crypt on GitHub)): One bearer key; no accounts.
- not this job: BitRun (Abhishek Khanra (IshekKhal on GitHub)): No key or accounts; one process serves one model.
- not this job: Docker Model Runner (Docker, Inc.): No credential or accounts; any client that reaches the socket can pull and run models.
- not this job: Foundry Local (Microsoft Corporation): The FAQ says it is single-user inference for one device and not a server stack; no concurrency or accounts.
- not this job: GPT4All (Nomic, Inc.): Localhost-only server with no authentication or accounts.
- not this job: HyperQwen (syv.ai (Copenhagen, CVR DK44671247)): VLLM_API_KEY only; no accounts.
- not this job: Jan (Menlo Research Pte Ltd): Desktop app for one user; the server has an optional single API key and no accounts.
- not this job: KoboldCpp (LostRuins (Concedo)): --multiuser queues several clients and --password is one shared key for text routes; no accounts or per-user view.
- not this job: Lemonade (Advanced Micro Devices, Inc.): One optional LEMONADE_API_KEY and an admin key; no accounts or groups.
- not this job: llama.cpp (ggml.ai, part of Hugging Face since 2026): Parallel slots and --api-key exist, but no accounts, groups or per-user view. The home-server account puts Open WebUI in front for the family.
- not this job: LM Studio (Element Labs, Inc.): One bearer token with Require Authentication; no accounts or groups. A desktop app and headless daemon for one owner.
- not this job: LocalGhost (LocalGhost.ai Ltd (company number 17213100, registered in England and Wales, incorporated 12 May 2026)): One box, one phone, no browser client or accounts.
- not this job: MLX LM (Apple Inc.): No API key option and --allowed-origins defaults to *; no accounts.
- not this job: Ollama (Ollama Inc.): No accounts, groups or credentials; the FAQ says to put an authenticating proxy in front. Ollama is the backend that Open WebUI deployments sit on, not the chat interface.
- not this job: PrivateGPT (PriBAI Technology Spain S.L.): One shared static secret when auth is on; no per-user accounts.
- not this job: screenpipe (Negentropy Labs, Inc. (dba Screenpipe)): Per-person recorder; Teams shares task configs, not a chat interface.
- not this job: TextGen (oobabooga): One --api-key at launch; no accounts or groups.
- not this job: Underdog (Conway Research): A personal assistant for one Mac over the owner's mail and calendar.
- not this job: vLLM (The Linux Foundation (vLLM is a PyTorch Foundation project)): --api-key only; no accounts or groups.

## What stops agents finishing

- API keys off by default and only an administrator can switch them on
- the model drops the system prompt after a few turns, so a persona or rule set does not hold
- a 500 MB desktop install or Docker stack is more than the person will set up

## Around it

Before: local-ai.fit-model-to-hardware. After: local-ai.chat-with-own-files, search.search-with-text.

## Where this comes from

Weight 13 of 100 in Local AI: 9 of 50 accounts, all Open WebUI or llama.cpp plus Open WebUI; two were deployments for hundreds of users

As data: https://www.anchorterminal.com/api/v1/use-cases/local-ai.household-or-team-chat.json, or the MCP tool get_use_case with id local-ai.household-or-team-chat. Draft, read 2026-11-12 from public docs and first-hand accounts, not yet reviewed by a person.

---

Query this catalogue directly over MCP: https://www.anchorterminal.com/mcp (Streamable HTTP, no key). How: https://www.anchorterminal.com/agents/#mcp
