BRI
Video intelligence service with a Streamlit operator surface, a FastAPI MCP server, SQLite durability, and multimodal tooling.
Proof ledger
- Classification
- Built artifact
- Repository
- Alexi5000/Bri
- Language
- Python
- Licence
- MIT
- Status
- Active development
- Tests
- 3 recorded gates
- Last verified
- 2026-08-03
What it is
BRI is a working example of exposing a multimodal pipeline through MCP rather than through a bespoke API. Whisper and YOLO tooling sit behind a FastAPI MCP server, state is durable in SQLite, and the operator surface is a separate Streamlit app.
Architecture
- FastAPI MCP server publishes the analysis tools with typed inputs.
- SQLite provides durable job and result state without an external database dependency.
- Separate Docker images for the MCP server and the UI so the tool surface can be deployed without the operator app.
- Pre-commit hooks and a Makefile standardise local runs.
Tests and gates
- Four workflows: ci, codeql, docs, and release.
- Pre-commit configuration and markdownlint run over the repository.
- GOVERNANCE.md, MAINTAINERS.md, and SECURITY.md define ownership and disclosure.
Limits
- Model quality is bounded by the underlying open models, not by the harness around them.
- SQLite durability suits single-node deployments; it is not a multi-writer data plane.
- The MCP surface is unauthenticated by default and expects to run behind a trusted boundary.
Source artifacts
Patterns applied
Acceptance tests for agent work →
Define done before the agent starts.
Evals that gate releases →
A golden suite in CI, not a spreadsheet.
MCP permission scoping →
Deny by default, declare every tool.
Observability for agent runs →
One trace ID from request to receipt.
Cost controls that actually stop spend →
Ceilings enforced before the call, not after the invoice.
Approval gates that suspend the run →
Nothing downstream executes before a human approves.
Related proof
ClawKeeper →
Governed Claude agent control plane. A deterministic policy engine runs before every LLM call, autonomous spend defaults to zero, and the audit log is enforced by the database.
TechTide Harness Kit →
Trust contracts for agent skills expressed as JSON Schema: risk tiers, tool permissions, deny-by-default egress, approval gates, and prompt-injection fixtures behind a wall of CI validation gates.
CipherClaw →
Deterministic multi-agent debugger with zero LLM calls in the analysis path: causal graphs, weighted root-cause scoring, and failure prediction.
Swarm 357 →
357 Claude agent roles across six business layers, with portable memory, enforced cost caps, a bash policy gate, and durable checkpoints.
This page is implementation proof only. Engagement scope and pricing live on TechTide AI.
AI Production Systems at TechTide AI→© 2026 Alex Cinovoj · TechTide AI · Columbus, OH