Questions, answered
The claims.
The mechanics.
The limits.
OpenWALDO makes strong claims about open source AI. Here are direct answers about what the project is, what its evidence establishes, and what remains your responsibility.
Start here
Understand
the project.
The short version: OpenWALDO is public infrastructure for identifying, verifying, selecting, and carrying the source material of AI into the models built from it.
Foundation
01What exactly is OpenWALDO?
OpenWALDO is an open training-data commons and the toolchain that makes it usable. The public index records corpus meaning and evidence in Git. Federated lookaside storage serves canonical, content-addressed Parquet objects. The WALDO command-line tool ingests, verifies, selects, exports, trains, and packages models while preserving provenance.
It is not a model vendor or a single frozen dataset. It is infrastructure for building models whose source material can be inspected.
02Why is open source AI important?
Transparency. Open source replaces unsupported assurances with evidence people can inspect: source material, license assertions, object identities, training plans, and recorded results. It does not guarantee that a model is safe or correct, but it makes meaningful independent scrutiny possible.
Community. Public inputs and open tooling let researchers, builders, creators, and users contribute improvements, find mistakes, challenge claims, and preserve work beyond the priorities of any one lab or vendor.
Scale. Shared infrastructure turns isolated efforts into a reusable foundation. Communities can divide the work of collecting, reviewing, hosting, verifying, and improving data instead of every organization rebuilding the same closed pipeline alone.
03Why aren’t open weights enough?
Weights are a valuable artifact, but they are the result of training—not the complete source needed to understand how the model came to exist. Without the training material, license assertions, object identity, configuration, and run history, users cannot fully inspect or reconstruct the process that produced them.
OpenWALDO’s position is direct: AI is not open source without the source.
04What is an AI Bill of Materials?
An OpenWALDO corpus BOM is an immutable, machine-readable record of one resolved data selection. It pins the index context, requested paths and license policy, manifest and source identities, resolved shards, exact counts, and hashes.
That record crosses boundaries. Corpus BOMs enter exports and training runs; origin, run, model, and release BOMs connect starting weights, execution, and published artifacts into one inspectable lineage.
05Does the training data live in Git?
No. Git governs the small, reviewable meaning: corpus descriptions, sources, asserted licenses, conversion identity, counts, and object references. Large canonical Parquet objects live in federated lookaside storage and are addressed by SHA-256.
This keeps review practical without hiding the bytes behind a private catalog.
Use the evidence
Select carefully.
Verify precisely.
WALDO separates recorded claims from machine checks so users can understand exactly what each piece of evidence does—and does not—establish.
Use and verification
06What does WALDO verify?
Metadata verification checks index structure and whether referenced objects are reachable at their declared sizes. Object verification downloads selected objects and proves their content hashes. A full audit also validates canonical records, identities, text hashes, duplicates, and aggregate totals.
These checks establish only what they examine. A hash proves identity and integrity; it does not prove license, quality, safety, or fitness for a model.
07How does OpenWALDO handle licenses?
The index records attributable license assertions and their scope so users can inspect and filter declared terms before selecting data. Assertions can be reviewed, challenged, and corrected through the public record.
OpenWALDO does not convert an assertion into a legal opinion. Commercial or other use depends on the selected material, its applicable terms, and the user’s own legal analysis.
08Can I use WALDO with existing models and training stacks?
Yes. Corpus selections can be exported as native Parquet or canonical JSONL with an offline-verifiable BOM, then consumed by another training system. WALDO can also manage training from a blank architecture or continue from a supported, revision-pinned open-weight model.
The managed model path intentionally supports a narrower set of architectures and tokenizers today. Export remains the general bridge to other stacks.
09Can a model be reproduced exactly?
WALDO can pin the exact corpus selection, starting-point identity, architecture, training plan, backend identity, environment observations, and output hashes. That makes the inputs and recorded process reproducible and differences diagnosable.
Bit-for-bit numerical reproduction is a separate claim and can depend on hardware, libraries, kernels, and backend determinism. WALDO records evidence; it does not promise equivalence the execution environment cannot provide.
Work in the open
Use it privately.
Improve it publicly.
The public commons is one deployment. The same formats and verification model can support local work, while public corrections and contributions strengthen the shared record.
Operation and participation
10Can WALDO run privately or offline?
Yes. WALDO supports local indexes, a retained local cache, and file-based lookaside storage. An organization can use the same content-addressed formats and verification workflow with internal material without publishing it to the public index.
Fully offline operation requires the selected index, objects, tools, and any model dependencies to be available locally.
11How do I contribute a corpus?
Bring local files or a reviewed acquisition recipe. WALDO probes the input, streams it into canonical Parquet, audits the result, publishes content-addressed objects, and prepares a compact Git overlay describing the large contribution’s reviewable meaning.
Normal repository review and DCO sign-off remain explicit. WALDO prepares the contribution; it does not silently commit or open a pull request.
12What does the Developer Certificate of Origin establish?
The DCO is a contributor attestation used across open-source projects. A
Signed-off-by trailer records that the contributor certifies they
created the submitted contribution—or otherwise have the right to submit it
under the project’s license—and understands that the contribution and sign-off
are public.
For WALDO, that puts an accountable identity behind the index change and its source and license assertions. It does not independently prove that those assertions are legally correct. Read how WALDO applies the DCO.
13How are mistakes, disputes, or removal requests handled?
Source and license statements are public, attributable claims—not untouchable conclusions. Questions and corrections should be raised against the index so evidence can be reviewed and the record can change visibly.
Object removal is deliberately explicit because an object may still be referenced by another index, historical revision, or exported BOM. WALDO does not guess that published data is unreferenced.
14What works today, and what is still planned?
Today WALDO covers canonical text ingestion, public indexing, verification and audit, BOM-backed export, model planning and training on supported backends, held-out evaluation, checkpointing, multiple release formats, regulatory disclosure projection, and optional signing.
The core path works end to end, and development is intentionally happening in public. Interfaces outside the documented compatibility contract may evolve as contributors test and improve them.
Multimodal ingestion, broader architecture and tokenizer compatibility, supervised and preference-training objectives, PyTorch generation, and multi-node orchestration remain deliberately pending. The current status and limits are maintained in the project documentation.
Go deeper
Inspect the record.
Question the claims.
The FAQ gives the short answer. The technical reference explains the current paths, while the community is where questions and challenges become better project decisions.