ClinicFlow AI
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| OLLAMA_HOST | No | Host and port for the local Ollama server used in real local model mode, e.g. 127.0.0.1:11435. | |
| DATABASE_URL | No | Database connection string for the SQLAlchemy domain layer (e.g. postgresql+psycopg://user:pass@host:port/db). The zero-service quickstart uses SQLite by default; PostgreSQL can be used instead. | |
| OLLAMA_NO_CLOUD | No | Set to 1 to disable Ollama cloud usage in real local model mode. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| available_appointmentsA | List synthetic clinic appointments; never book from model output alone. |
| propose_appointmentB | Validate an available slot and return a proposal needing human confirmation. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 2 tools
The two tools are clearly separable: one is a read-only listing of slots, the other a validation/proposal action that produces a human-confirmation artifact. There is mild overlap since both operate on appointment slots, but a caller can reliably pick the right one.
Naming styles are mixed: 'available_appointments' is an adjective+noun phrase while 'propose_appointment' follows a verb_noun pattern. Both remain readable and snake_case is consistent, but there is no single predictable convention.
Two tools is thin even accounting for the deliberate human-in-the-loop safety constraint. The pair covers discovery and proposal, but the surface feels minimal for a clinic booking domain.
Only listing and proposing exist; there is no way to confirm a proposal, cancel, reschedule, or fetch details of a specific appointment. Some of this is intentionally offloaded to a human, but an agent will hit dead ends for common follow-up operations.