A good omni-directional microphone is a good start, if you can't have a dedicated mic for each person. I've used a Snowball switched to its omnidirectional setting before, and it's good enough for documenting a session with good-enough sound quality that you can hear everything. (For the quality of a good postcast, a single mic won't cut it and you'll want dedicated mics, and might consider a seating arrangement that prioritises recording quality and sound separation. That's beyond my experience, but sounds like it's beyond what you need anyway.)
As for software, I've used Audacity to record and post-process play session audio, for personal documentation and sharing purposes. It's well-suited to the job as it gracefully handles the very large file sizes that you'll get when recording hours of audio. As an editor and post-processing tool it is also quite powerful, and it exports to various compressed formats for fize sizes that are more reasonable for sharing.
A one-shot session of Dungeon World is pretty trivial, actually. So far, all games of DW I've played have been one-shots. All you need to prepare as DM is a short adventure, typically a dungeon. You don't need to deal with Fronts at all, and setup is the normal quick character-creation process of throwing the playbooks at the players and asking them to fill in the blanks. You just run the game like it's a First Session (p. 177), except that there won't be a second session so you don't have to do the After the First Session exercises. The checklist of goals on page 180 is what you should focus on.
To complete an adventure in a single session, you need either an adventure that you've got a really good handle on how to pace to fit, or design (or borrow) one that has an open-ended, players-decided goal. A dungeon that they can choose to back out of at any time is one example of such an open-ended, player-controlled adventure. Completing an adventure isn't really necessary, but something you can try if you want. The players will have fun regardless of whether the adventure concludes. The worst that can happen is that they ask to play again to find out what happens next, right?
My first experience was with Tony Dowler's Purple Worm Graveyard: we started at the doors of the dungeon, and the objective was to return home with wealth or die trying. As it was a convention game, we could choose at any time to back out and take "home" whatever wealth we'd found. (Not that we couldn't have done the same in a non-convention game, but it was helpful for the DM to say we could do that explicitly so that we could have a satisfying conclusion within the time allotted.) We didn't explore the whole dungeon (and it's not even a very large one), but we had lots of fun, enjoyed the open-ended options of the game just fine, and escaped the dungeon with a satisfying amount of loot and intangible discoveries. In a two-hour session we got through seven chambers, so a budget of 15–30 minutes per room is probably a reasonable rule of thumb when you're drawing a dungeon for a one-shot of DW.
My second experience playing DW (also at a con) was an investigation adventure that involved a town, a mayor with secrets, and a church. We didn't finish the adventure in the time available, but we had a lot of fun and thoroughly enjoyed the open-endedness and player agency that DW offers. I would have liked to finish uncovering what was going on with that cult, but it didn't detract from enjoying the play process in the least.
Best Answer
I've One-upped all of the answers that are already here.
You have two issues, how to create an adventure quickly, and how to keep the workload down.
Pregenerated characters are a good idea.. BUT you may want to get the PCs to build them just to save your own workload. To compromise- give the PCs some guidelines: "you are all playing dwarves" or "Arcane power source only" or "Just Essentials PCs" is a good way to get everything focused from the start. You might say "two of you are strikers, one of you is a leader, and one of you is a defender. Your choice which is which. This guidenace gets the players started about what kind of character they should make if you decide to go that way.
Both the Delves, and the Encounter group type deals are also good calls. If you have access to the Adventure Builder/DDI, it's also pretty easy to shoot out some printouts of monster blocks and recombine them into encounters according to the formula that an encounter budget is equal to
So if you print out a single kobold slinger, kobold dragonshield, kobold miner (minion) and kobold wyrmpriest- you can turn that into 4 (or 40) different encounters just by altering the mixup.
This is a standard Dungeoneering Adventure Plan. There's nothing amazing about this, but it does work, especially and as long as you remember that the "story" is what players and their characters do, not anything you plan for.
Grab a big stack of dungeon tiles or just creative with a battlemap and markers.
Improvise a reason for adventurers to take on a monster-lair. Even just "there's treasure there" or "they are kidnapping people.."
As the players explore- describe, improvise the lair room by room. It could be 3 rooms or 30 or 300. But there are only 2-4 encounters you need to run. Be sure to improvise terrain and atmosphere! You might even draw dungeon tiles randomly from a deck.
Whenever you get to an encounter, drop in an improvised mix of the monsters you chose that somehow approximates the encounter formula. Be sure to improvise terrain and atmosphere again!
Plan 3 or 4 treasure drops. Some of the treasure might be unguarded, trapped, hidden, or inaccessible (for example, a ring at the bottom of a deep well, or a chest behind a closed porticullis gate).
Consider putting at least one skill challenge (perhaps a complicated puzzle or trap) and at least one NPC encounter (in a dungeon perhaps this could be an encounter with prisoners, a ghost, a fellow adventurer, a crazy outcast kobold..or anything else).
No story planned, improvise everything. Let the PCs explore the characters, and do whatever they like. Go with the flow, play off what the PCs do.
Example of an Improv 4e Adventure:
If you have 4 level 1 characters, that's about 400xp to work with.
early in the adventure you might use 2 dragonshields (thats 250xp) and 2 slingers (200xp) that have set themselves up with some barricades and such. Guards.
Later on, after exploring several rooms, the kobolds ambush from a room with overhanging ledges. Use 2 more slingers, a wyrmpriest and 4 minions.
Later on the PCs encounter a group of enslaved miners, whom they rescue. And the miners tell the PCs about a dragon-idol the kobolds are forcing them to excavate. Roleplay roleplay. Invent NPCs. Do it on the spot if necessary.
Later on the PCs might come across the dragon idol itself, which is magically unstable, intelligent and able to collapse part of the caverns unless the pCs intervene via religion, history, arcana, and theivery- to calm it/remove it's magical residuum core.etc. Skill challenge.
Improvise some tracks or trails that lead through the kobold warrens to a final lair.
Wrap up the final encounter with a wyrmpriest, a dragonshield and 10 minions and the rest of the treasure.
To make it less of a railroad, generate the encounters beforehand (and create 2x or 3x as many) and create a map and matrix. That's really not that much more work, especially in 4e.