Run legal work on the top open-source models at a fraction of what closed assistants cost — up to 1M tokens of context, 44 legal skills, legal agents, and a provenance ledger that labels every claim. More work, more tokens, less spend.
Built for solicitors, barristers, in-house counsel and legal ops
DeepSeek V4.1 Flash by default, plus Qwen, Llama 4, Mistral, GLM, Kimi and Gemma — up to 1M tokens of context.
Contract review, citation verification, research memos, intake, deadlines, governance and more — loaded on demand.
Contract Reviewer, Citation Verifier, Memo Drafter and Client Intake — each with the method built in.
Every claim labelled [AUTHORITY], [REASONING] or [NOT IN SOURCE], traceable to a pin cite.
Upload files, search the web, and keep matter context across sessions.
Your own account and storage; no training on your data; human sign-off always.
We route your work to the best open-source models through one account. On typical drafting and review, that can be a small fraction of the per-token cost of closed frontier assistants — often up to 100× cheaper.
| Model | Context | Typical input / 1M tokens |
|---|---|---|
| DeepSeek V4.1 Flash (default) | 1,000,000 | ~$0.02 |
| Qwen 3.8 Max | 1,000,000 | ~$0.10 |
| Mistral Large 4 | 1,000,000 | ~$0.30 |
| Closed frontier assistant (illustrative) | ~200k | ~$3–15 |
Illustrative rates, October 2026; provider prices vary. You pay for what you use — no per-seat model licence.
Contract Reviewer, Citation Verifier, Memo Drafter and Intake agents, plus 44 installable legal skills.
Upload files and chat with them (RAG); keep matter context across sessions.
Search the web and trace claims to a source — with [NOT IN SOURCE] when it can't.
Admins set the default and allowed models; users pick from a curated list.
Firm plans can use their own OpenRouter/provider keys, or ours.
Team workspace, roles, single sign-on and audit for firm deployments.
Watch a contract go in, red flags come out, and a verified memo with its provenance ledger.
Demo video coming soon
We deploy and manage an open-source AI stack in your own environment — your models, your data, your policies, with SSO, audit and support. Enterprise scoping and pricing.
Per-user pricing. Cancel anytime. Or pay as you go with credit packs.
For one lawyer testing AI on real work.
For a lawyer who drafts every day.
For teams and in-house departments.
No subscription. Buy credits and use them across models and agents.
Self-hosted, isolated accounts, encrypted credentials and private file storage. Agents practise and score; they never issue the verification claim — a named lawyer signs off.
No. It is an information and drafting tool. Every output is labelled and a qualified lawyer must review and sign off before it is relied on.
DeepSeek V4.1 Flash by default, plus Qwen, Llama 4, Mistral, GLM, Kimi and Gemma — up to 1M tokens of context.
In your own account on infrastructure we control, with encrypted credentials and private document storage. You can bring your own model keys on Firm plans.
Yes — buy credit packs with no subscription and spend them across models and agents.