Core + Trust · Forecast-Driven Workflows

Tollama AI

Forecast-driven decisions
need evidence first.

Tollama Core preprocesses irregular series, runs forecasts, benchmarks models, and routes future requests from evidence. Tollama Trust decides whether those outputs are reliable enough to act on in production workflows.

Get Started View on GitHub ↗
Forecasting is
fragmented.

Teams still stitch together preprocessing, model serving, one-off benchmarks, and trust decisions by hand. There is rarely one clear path from irregular series to an auditable production action.

01 / Serving
Fragmented runtimes.

Every TSFM ships its own install, APIs, and deployment patterns. Serving multiple models means managing multiple stacks with conflicting dependencies.

02 / Evaluation
Unfair evaluation.

One-off notebooks with random splits, inconsistent metrics, and no reproducibility. No standardized way to compare models on your data.

03 / Decision Risk
Low-confidence outputs still reach action.

Even when a forecast exists, teams often lack a consistent way to score trust, trigger review, or write an audit trail before acting on it.

// Products
Two products.
One forecast-decision stack.

Tollama Core is the OSS front door for preprocess -> forecast -> benchmark -> route. Its first concrete solution is benchmark-backed hourly demand forecasting for operations teams. Tollama Trust decides whether those outputs are ready for production action.

Production
Tollama Core
The OSS front door for benchmark-backed hourly demand forecasting.

Local-first time-series forecasting core for irregular-series preprocessing, unified forecasting, benchmark artifacts, and benchmark-backed routing defaults, with a checked-in concrete-solution path.

Core OSS Hourly demand demo Thin artifact bundle Route from evidence
Learn more →
🛡️
Production
Tollama Trust
The decision-trust layer for forecast workflows.

Scores whether outputs are reliable enough to act on, applies policy gates such as allow, review, and block, and writes auditable trust records.

Trust scores Policy gates Audit trail
Learn more →
// Supporting Assets
Supporting layers.
Not separate products.

These assets strengthen the Core + Trust story, but they are not the primary product entry points.

🔬 tollama-eval Beta

The richer benchmark layer behind Core artifacts. Use it when Core needs broader model sweeps, richer reports, and deeper evaluation evidence.

View product →
🧠 spline-lstm Production

Preprocessing lineage and spline differentiation for irregular-series cleanup, interpolation, and window preparation.

View product →
📊 Market Calibration Agent Production

Hero wedge for trust-aware market workflows. Calibration analysis and market signal trust scoring connect Core forecasts to real decision contexts.

View product →
// Roadmap
What's next.

The next priority is not more breadth. It is making Core easier to demo, Trust easier to attach, and benchmark evidence easier to operationalize.

UPCOMING
Concrete Solution Pack

Keep the checked-in hourly-demand input, expected-output bundle, and smoke-gated demo path aligned as the Core front door evolves.

UPCOMING
Continuous Evaluation

Scheduled benchmark runs, regression checks, and release gating for forecast systems.

UPCOMING
Trust Attach-On

A smaller, clearer Trust surface that reads as a decision layer on top of Core evidence.

UPCOMING
Hero Market Demo

A visible end-to-end demo that connects preprocessing quality, forecast output, trust score, and gate result.

UPCOMING
Decision-Ready Output

Keep operator summaries, routing rationale, and Trust handoff IDs visible wherever Core artifacts are shown.

// Get Started

Start with Core.
Attach Trust when ready.

Build forecast-driven workflows from benchmark-backed evidence first, then add decision trust where action risk matters.

View on GitHub ↗ Explore Products →